def compute_query_projection(
what: Union[str, ut.Matrix] = "__x__",
*,
adata: AnnData,
qdata: AnnData,
weights: ut.Matrix,
atlas_total_umis: Optional[ut.Vector] = None,
query_total_umis: Optional[ut.Vector] = None,
) -> None:
"""
Compute the projected image of the query on the atlas.
**Input**
Annotated query ``qdata`` and atlas ``adata``, where the observations are cells and the variables are genes, where
``what`` is a per-variable-per-observation matrix or the name of a per-variable-per-observation annotation
containing such a matrix.
The ``weights`` of the projection where each row is a query metacell, each column is an atlas metacell, and the
value is the weight of the atlas cell for projecting the metacell, such that the sum of weights in each row
is one.
**Returns**
In addition, sets the following annotations in ``qdata``:
Observation (Cell) Annotations
``projection``
The number of UMIs of each gene in the projected image of the query to the metacell, if the total number of
UMIs in the projection is equal to the total number of UMIs in the query metacell.
**Computation Parameters**
1. Compute the fraction of each gene in the atlas and the query based on the total UMIs, unless ``atlas_total_umis``
and/or ``query_total_umis`` are specified.
2. Compute the projected image of each query metacell on the atlas using the weights.
3. Convert this image to UMIs count based on the total UMIs of each metacell. Note that if overriding the total
atlas or query UMIs, this means that the result need not sum to this total.
"""
assert np.all(adata.var_names == qdata.var_names)
atlas_umis = ut.get_vo_proper(adata, what, layout="row_major")
query_umis = ut.get_vo_proper(qdata, what, layout="row_major")
if atlas_total_umis is None:
atlas_total_umis = ut.sum_per(atlas_umis, per="row")
atlas_total_umis = ut.to_numpy_vector(atlas_total_umis)
if query_total_umis is None:
query_total_umis = ut.sum_per(query_umis, per="row")
query_total_umis = ut.to_numpy_vector(query_total_umis)
atlas_fractions = ut.to_numpy_matrix(ut.fraction_by(atlas_umis, by="row", sums=atlas_total_umis))
projected_fractions = weights @ atlas_fractions # type: ignore
projected_umis = ut.scale_by(projected_fractions, scale=query_total_umis, by="row")
ut.set_vo_data(qdata, "projected", projected_umis)
def _pick_candidates(
*,
adata_of_all_genes_of_all_cells: AnnData,
what: Union[str, ut.Matrix] = "__x__",
max_gene_cell_fraction: float,
min_gene_maximum: int,
min_genes_of_modules: int,
allowed_genes_mask: ut.NumpyVector,
) -> Optional[Tuple[AnnData, ut.NumpyVector]]:
data = ut.get_vo_proper(adata_of_all_genes_of_all_cells,
what,
layout="column_major")
nnz_cells_of_genes = ut.nnz_per(data, per="column")
nnz_cell_fraction_of_genes = nnz_cells_of_genes / adata_of_all_genes_of_all_cells.n_obs
nnz_cell_fraction_mask_of_genes = nnz_cell_fraction_of_genes <= max_gene_cell_fraction
max_umis_of_genes = ut.max_per(data, per="column")
max_umis_mask_of_genes = max_umis_of_genes >= min_gene_maximum
candidates_mask_of_genes = max_umis_mask_of_genes & nnz_cell_fraction_mask_of_genes & allowed_genes_mask
ut.log_calc("candidate_genes", candidates_mask_of_genes)
candidate_genes_indices = np.where(candidates_mask_of_genes)[0]
candidate_genes_count = candidate_genes_indices.size
if candidate_genes_count < min_genes_of_modules:
return None
candidate_data = ut.slice(adata_of_all_genes_of_all_cells,
name=".candidate_genes",
vars=candidate_genes_indices,
top_level=False)
return candidate_data, candidate_genes_indices
def find_high_relative_variance_genes(
adata: AnnData,
what: Union[str, ut.Matrix] = "__x__",
*,
min_gene_relative_variance: float = pr.significant_gene_relative_variance,
window_size: int = pr.relative_variance_window_size,
inplace: bool = True,
) -> Optional[ut.PandasSeries]:
"""
Find genes which have high relative variance of ``what`` (default: {what}) data.
The relative variance measures the variance / mean of each gene relative to the other genes with
a similar level of expression. See
:py:func:`metacells.utilities.computation.relative_variance_per` for details.
Genes with a high relative variance are good candidates for being selected as "feature genes",
that is, be used to compute the similarity between cells. Using the relative variance
compensates for the bias for selecting higher-expression genes, whose normalized variance can to
be larger due to random noise alone.
**Input**
Annotated ``adata``, where the observations are cells and the variables are genes, where
``what`` is a per-variable-per-observation matrix or the name of a per-variable-per-observation
annotation containing such a matrix.
**Returns**
Variable (Gene) Annotations
``high_relative_variance_gene``
A boolean mask indicating whether each gene was found to have a high relative
variance.
If ``inplace`` (default: {inplace}), this is written to the data, and the function returns
``None``. Otherwise this is returned as a pandas series (indexed by the variable names).
**Computation Parameters**
1. Use :py:func:`metacells.utilities.computation.relative_variance_per` to get the relative
variance of each gene.
2. Select the genes whose relative variance is at least
``min_gene_relative_variance`` (default: {min_gene_relative_variance}).
"""
data = ut.get_vo_proper(adata, what, layout="column_major")
relative_variance_of_genes = ut.relative_variance_per(data, per="column", window_size=window_size)
genes_mask = relative_variance_of_genes >= min_gene_relative_variance
if inplace:
ut.set_v_data(adata, "high_relative_variance_gene", genes_mask)
return None
ut.log_return("high_relative_variance_genes", genes_mask)
return ut.to_pandas_series(genes_mask, index=adata.var_names)
def find_high_normalized_variance_genes(
adata: AnnData,
what: Union[str, ut.Matrix] = "__x__",
*,
min_gene_normalized_variance: float = pr.significant_gene_normalized_variance,
inplace: bool = True,
) -> Optional[ut.PandasSeries]:
"""
Find genes which have high normalized variance of ``what`` (default: {what}) data.
The normalized variance measures the variance / mean of each gene. See
:py:func:`metacells.utilities.computation.normalized_variance_per` for details.
Genes with a high normalized variance are "noisy", that is, have significantly different
expression level in different cells.
**Input**
Annotated ``adata``, where the observations are cells and the variables are genes, where
``what`` is a per-variable-per-observation matrix or the name of a per-variable-per-observation
annotation containing such a matrix.
**Returns**
Variable (Gene) Annotations
``high_normalized_variance_gene``
A boolean mask indicating whether each gene was found to have a high normalized
variance.
If ``inplace`` (default: {inplace}), this is written to the data, and the function returns
``None``. Otherwise this is returned as a pandas series (indexed by the variable names).
**Computation Parameters**
1. Use :py:func:`metacells.utilities.computation.normalized_variance_per` to get the normalized
variance of each gene.
2. Select the genes whose normalized variance is at least
``min_gene_normalized_variance`` (default: {min_gene_normalized_variance}).
"""
data = ut.get_vo_proper(adata, what, layout="column_major")
normalized_variance_of_genes = ut.normalized_variance_per(data, per="column")
genes_mask = normalized_variance_of_genes >= min_gene_normalized_variance
if inplace:
ut.set_v_data(adata, "high_normalized_variance_gene", genes_mask)
return None
ut.log_return("high_normalized_variance_genes", genes_mask)
return ut.to_pandas_series(genes_mask, index=adata.var_names)
def find_biased_genes(
adata: AnnData,
*,
max_projection_fold_factor: float = pr.project_max_projection_fold_factor,
min_metacells_fraction: float = pr.biased_min_metacells_fraction,
abs_folds: bool = pr.project_abs_folds,
to_property_name: str = "biased_gene",
) -> None:
"""
Find genes that have a strong bias in the query compared to the atlas.
**Input**
Annotated query ``adata`` where the observations are cells and the variables are genes, where ``what`` is a
per-variable-per-observation matrix or the name of a per-variable-per-observation annotation containing such a
matrix.
This should contain a ``projected_fold`` per-variable-per-observation matrix with the fold factor between each query
metacell and its projected image on the atlas.
**Returns**
Sets the following annotations in ``adata``:
Variable (Gene) Annotations
``biased_gene`` (or ``to_property_name``):
A boolean mask indicating whether the gene has a strong bias in the query compared to the atlas.
**Computation Parameters**
1. Count for each such gene the number of query metacells for which the ``projected_fold`` is above
``max_projection_fold_factor``. If ``abs_folds`` (default: {abs_folds}), consider the absolute fold factor.
2. Mark the gene as biased if either count is at least a ``min_metacells_fraction`` (default:
{min_metacells_fraction}) of the metacells.
"""
assert max_projection_fold_factor >= 0
assert 0 <= min_metacells_fraction <= 1
projected_fold = ut.get_vo_proper(adata, "projected_fold", layout="column_major")
if abs_folds:
projected_fold = np.abs(projected_fold) # type: ignore
high_projection_folds = ut.to_numpy_matrix(projected_fold > max_projection_fold_factor) # type: ignore
ut.log_calc("high_projection_folds", high_projection_folds)
count_of_genes = ut.sum_per(high_projection_folds, per="column")
min_count = adata.n_obs * min_metacells_fraction
mask_of_genes = count_of_genes >= min_count
ut.set_v_data(adata, to_property_name, mask_of_genes)
def find_high_topN_genes( # pylint: disable=invalid-name
adata: AnnData,
what: Union[str, ut.Matrix] = "__x__",
*,
topN: int, # pylint: disable=invalid-name
min_gene_topN: int, # pylint: disable=invalid-name
inplace: bool = True,
) -> Optional[ut.PandasSeries]:
"""
Find genes which have high total top-Nth value of ``what`` (default: {what}) data.
This should typically only be applied to downsampled data to ensure that variance in sampling
depth does not affect the result.
Genes with too-low expression are typically excluded from computations. In particular,
genes may have all-zero expression, in which case including them just slows the
computations (and triggers numeric edge cases).
**Input**
Annotated ``adata``, where the observations are cells and the variables are genes, where
``what`` is a per-variable-per-observation matrix or the name of a per-variable-per-observation
annotation containing such a matrix.
**Returns**
Variable (Gene) Annotations
``high_top<topN>_gene``
A boolean mask indicating whether each gene was found to have a high top-Nth value.
If ``inplace`` (default: {inplace}), this is written to the data, and the function returns
``None``. Otherwise this is returned as a pandas series (indexed by the variable names).
**Computation Parameters**
1. Use :py:func:`metacells.utilities.computation.top_per` to get the top-Nth UMIs of each gene.
2. Select the genes whose fraction is at least ``min_gene_topN``.
"""
data_of_genes = ut.get_vo_proper(adata, what, layout="column_major")
rank = max(adata.n_obs - topN - 1, 1)
topN_of_genes = ut.rank_per(data_of_genes, per="column", rank=rank) # pylint: disable=invalid-name
genes_mask = topN_of_genes >= min_gene_topN
if inplace:
ut.set_v_data(adata, f"high_top{topN}_gene", genes_mask)
return None
ut.log_return(f"high_top{topN}_genes", genes_mask)
return ut.to_pandas_series(genes_mask, index=adata.var_names)
def find_high_fraction_genes(
adata: AnnData,
what: Union[str, ut.Matrix] = "__x__",
*,
min_gene_fraction: float = pr.significant_gene_fraction,
inplace: bool = True,
) -> Optional[ut.PandasSeries]:
"""
Find genes which have high fraction of the total ``what`` (default: {what}) data of the cells.
Genes with too-low expression are typically excluded from computations. In particular,
genes may have all-zero expression, in which case including them just slows the
computations (and triggers numeric edge cases).
**Input**
Annotated ``adata``, where the observations are cells and the variables are genes, where
``what`` is a per-variable-per-observation matrix or the name of a per-variable-per-observation
annotation containing such a matrix.
**Returns**
Variable (Gene) Annotations
``high_fraction_gene``
A boolean mask indicating whether each gene was found to have a high normalized
variance.
If ``inplace`` (default: {inplace}), this is written to the data, and the function returns
``None``. Otherwise this is returned as a pandas series (indexed by the variable names).
**Computation Parameters**
1. Use :py:func:`metacells.utilities.computation.fraction_per` to get the fraction of each gene.
2. Select the genes whose fraction is at least ``min_gene_fraction`` (default:
{min_gene_fraction}).
"""
data = ut.get_vo_proper(adata, what, layout="column_major")
fraction_of_genes = ut.fraction_per(data, per="column")
genes_mask = fraction_of_genes >= min_gene_fraction
if inplace:
ut.set_v_data(adata, "high_fraction_gene", genes_mask)
return None
ut.log_return("high_fraction_genes", genes_mask)
return ut.to_pandas_series(genes_mask, index=adata.var_names)
def compute_similar_query_metacells(
adata: AnnData,
max_projection_fold_factor: float = pr.project_max_projection_fold_factor,
abs_folds: bool = pr.project_abs_folds,
) -> None:
"""
Mark query metacells that are similar to their projection on the atlas.
This does not guarantee the query metacell is "the same as" its projection on the atlas; rather, it means the two
are sufficiently similar that one can be "reasonably confident" in applying atlas metadata to the query metacell
based on the projection, which is a much lower bar.
**Input**
Annotated query ``adata``, where the observations are cells and the variables are genes, where ``what`` is a
per-variable-per-observation matrix or the name of a per-variable-per-observation annotation containing such a
matrix.
The data should contain per-observation-per-variable annotations ``projected_fold`` with the significant projection
folds factors, as computed by :py:func:`compute_significant_projected_fold_factors`.
**Returns**
Sets the following in ``adata``:
Per-Observation (Cell) Annotations
``similar``
A boolean mask indicating the query metacell is similar to its projection in the atlas.
**Computation Parameters**
1. Mark as dissimilar any query metacells which have even one gene whose projection fold is above
``max_projection_fold_factor``.
"""
assert max_projection_fold_factor >= 0
projected_folds = ut.get_vo_proper(adata,
"projected_fold",
layout="row_major")
if abs_folds:
projected_folds = np.abs(projected_folds) # type: ignore
high_folds = projected_folds > max_projection_fold_factor # type: ignore
high_folds_per_metacell = ut.sum_per(high_folds, per="row") # type: ignore
similar_mask = high_folds_per_metacell == 0
ut.set_o_data(adata, "similar", similar_mask)
def find_distinct_genes(
adata: AnnData,
what: Union[str, ut.Matrix] = "distinct_fold",
*,
distinct_genes_count: int = pr.distinct_genes_count,
distinct_abs_folds: bool = pr.distinct_abs_folds,
inplace: bool = True,
) -> Optional[Tuple[ut.PandasFrame, ut.PandasFrame]]:
"""
Find for each observation (cell) the genes in which its ``what`` (default: {what}) value is most
distinct from the general population. This is typically applied to the metacells data rather
than to the cells data.
**Input**
Annotated ``adata``, where the observations are (mata)cells and the variables are genes,
including a per-observation-per-variable annotated folds data, {what}), e.g. as computed by
:py:func:`compute_distinct_folds`.
**Returns**
Observation-Any (Cell) Annotations
``cell_distinct_gene_indices``
For each cell, the indices of its top ``distinct_genes_count`` genes.
``cell_distinct_gene_folds``
For each cell, the fold factor of its top ``distinct_genes_count``.
If ``inplace`` (default: {inplace}), this is written to the data, and the function returns
``None``. Otherwise this is returned as two pandas frames (indexed by the observation and
distinct gene rank).
**Computation Parameters**
1. Fetch the previously computed per-observation-per-variable ``what`` data.
2. Keep the ``distinct_genes_count`` (default: {distinct_genes_count}) top fold factors. If ``distinct_abs_folds``
(default: ``distinct_abs_folds``), keep the top absolute fold factors.
"""
assert 0 < distinct_genes_count < adata.n_vars
distinct_gene_indices = np.empty((adata.n_obs, distinct_genes_count),
dtype="int32")
distinct_gene_folds = np.empty((adata.n_obs, distinct_genes_count),
dtype="float32")
fold_in_cells = ut.mustbe_numpy_matrix(
ut.get_vo_proper(adata, what, layout="row_major"))
extension_name = f"top_distinct_{fold_in_cells.dtype}_t"
extension = getattr(xt, extension_name)
extension(distinct_gene_indices, distinct_gene_folds, fold_in_cells,
distinct_abs_folds)
if inplace:
ut.set_oa_data(adata, "cell_distinct_gene_indices",
distinct_gene_indices)
ut.set_oa_data(adata, "cell_distinct_gene_folds", distinct_gene_folds)
return None
return (
ut.to_pandas_frame(distinct_gene_indices, index=adata.obs_names),
ut.to_pandas_frame(distinct_gene_folds, index=adata.obs_names),
)
def find_deviant_cells(
adata: AnnData,
what: Union[str, ut.Matrix] = "__x__",
*,
candidates: Union[str, ut.Vector] = "candidate",
min_gene_fold_factor: float = pr.deviants_min_gene_fold_factor,
abs_folds: bool = pr.deviants_abs_folds,
max_gene_fraction: Optional[float] = pr.deviants_max_gene_fraction,
max_cell_fraction: Optional[float] = pr.deviants_max_cell_fraction,
inplace: bool = True,
) -> Optional[Tuple[ut.PandasSeries, ut.PandasSeries]]:
"""
Find cells which are have significantly different gene expression from the metacells they are
belong to based on ``what`` (default: {what}) data.
**Input**
Annotated ``adata``, where the observations are cells and the variables are genes, where
``what`` is a per-variable-per-observation matrix or the name of a per-variable-per-observation
annotation containing such a matrix.
**Returns**
Observation (Cell) Annotations
``cell_deviant_votes``
The number of genes that were the reason the cell was marked as deviant (if zero, the
cell is not deviant).
Variable (Gene) Annotations
``gene_deviant_votes``
The number of cells each gene marked as deviant (if zero, the gene did not mark any cell
as deviant).
If ``inplace`` (default: {inplace}), this is written to the data, and the function returns
``None``. Otherwise this is returned as two pandas series (indexed by the observation and
variable names).
**Computation Parameters**
Intuitively, we first select some fraction of the genes which were least predictable compared to
the mean expression in the candidate metacells. We then mark as deviants some fraction of the
cells whose expression of these genes was least predictable compared to the mean expression in
the candidate metacells. Operationally:
1. Compute for each candidate metacell the mean fraction of the UMIs expressed by each gene.
Scale this by each cell's total UMIs to compute the expected number of UMIs for each cell.
Compute the fold factor log2((actual UMIs + 1) / (expected UMIs + 1)) for each gene for each
cell.
2. Ignore all fold factors less than the ``min_gene_fold_factor`` (default: {min_gene_fold_factor}). If
``abs_folds`` (default: {abs_folds}), consider the absolute fold factors. Count the number of genes which have a
fold factor above this minimum in at least one cell. If the fraction of such genes is above ``max_gene_fraction``
(default: {max_gene_fraction}), then raise the minimal gene fold factor such that at most this fraction of genes
remain.
3. For each remaining gene, rank all the cells where it is expressed above the min fold
factor. Give an artificial maximum rank to all cells with fold factor 0, that is, below the
minimum.
4. For each cell, compute the minimal rank it has in any of these genes. That is, if a cell has
a rank of 1, it means that it has at least one gene whose expression fold factor is the worst
(highest) across all cells (and is also above the minimum).
5. Select as deviants all cells whose minimal rank is below the artificial maximum rank, that
is, which contain at least one gene whose expression fold factor is high relative to the rest
of the cells. If the fraction of such cells is higher than ``max_cell_fraction`` (default:
{max_cell_fraction}), reduce the maximal rank such that at most this fraction of cells are
selected as deviants.
"""
if max_gene_fraction is None:
max_gene_fraction = 1
if max_cell_fraction is None:
max_cell_fraction = 1
assert min_gene_fold_factor > 0
assert 0 < max_gene_fraction < 1
assert 0 < max_cell_fraction < 1
cells_count, genes_count = adata.shape
assert cells_count > 0
candidate_of_cells = ut.get_o_numpy(adata, candidates, formatter=ut.groups_description)
totals_of_cells = ut.get_o_numpy(adata, what, sum=True)
assert totals_of_cells.size == cells_count
data = ut.get_vo_proper(adata, what, layout="row_major")
list_of_fold_factors, list_of_cell_index_of_rows = _collect_fold_factors(
data=data,
candidate_of_cells=candidate_of_cells,
totals_of_cells=totals_of_cells,
min_gene_fold_factor=min_gene_fold_factor,
abs_folds=abs_folds,
)
fold_factors = _construct_fold_factors(cells_count, list_of_fold_factors, list_of_cell_index_of_rows)
if fold_factors is None:
votes_of_deviant_cells = np.zeros(adata.n_obs, dtype="int32")
votes_of_deviant_genes = np.zeros(adata.n_vars, dtype="int32")
else:
deviant_gene_indices = _filter_genes(
cells_count=cells_count,
genes_count=genes_count,
fold_factors=fold_factors,
min_gene_fold_factor=min_gene_fold_factor,
max_gene_fraction=max_gene_fraction,
)
deviant_genes_fold_ranks = _fold_ranks(
cells_count=cells_count, fold_factors=fold_factors, deviant_gene_indices=deviant_gene_indices
)
votes_of_deviant_cells, votes_of_deviant_genes = _filter_cells(
cells_count=cells_count,
genes_count=genes_count,
deviant_genes_fold_ranks=deviant_genes_fold_ranks,
deviant_gene_indices=deviant_gene_indices,
max_cell_fraction=max_cell_fraction,
)
if inplace:
ut.set_v_data(adata, "gene_deviant_votes", votes_of_deviant_genes, formatter=ut.mask_description)
ut.set_o_data(adata, "cell_deviant_votes", votes_of_deviant_cells, formatter=ut.mask_description)
return None
ut.log_return("gene_deviant_votes", votes_of_deviant_genes, formatter=ut.mask_description)
ut.log_return("cell_deviant_votes", votes_of_deviant_cells, formatter=ut.mask_description)
return (
ut.to_pandas_series(votes_of_deviant_cells, index=adata.obs_names),
ut.to_pandas_series(votes_of_deviant_genes, index=adata.var_names),
)
def dissolve_metacells(
adata: AnnData,
what: Union[str, ut.Matrix] = "__x__",
*,
candidates: Union[str, ut.Vector] = "candidate",
deviants: Optional[Union[str, ut.Vector]] = "cell_deviant_votes",
target_metacell_size: float = pr.target_metacell_size,
cell_sizes: Optional[Union[str, ut.Vector]] = pr.dissolve_cell_sizes,
min_metacell_cells: int = pr.dissolve_min_metacell_cells,
min_robust_size_factor: Optional[float] = pr.
dissolve_min_robust_size_factor,
min_convincing_size_factor: Optional[float] = pr.
dissolve_min_convincing_size_factor,
min_convincing_gene_fold_factor: float = pr.
dissolve_min_convincing_gene_fold_factor,
abs_folds: bool = pr.dissolve_abs_folds,
inplace: bool = True,
) -> Optional[ut.PandasFrame]:
"""
Dissolve too-small metacells based on ``what`` (default: {what}) data.
**Input**
Annotated ``adata``, where the observations are cells and the variables are genes, where
``what`` is a per-variable-per-observation matrix or the name of a per-variable-per-observation
annotation containing such a matrix.
**Returns**
Observation (Cell) Annotations
``metacell``
The integer index of the metacell each cell belongs to. The metacells are in no
particular order. Cells with no metacell assignment are given a metacell index of
``-1``.
``dissolved``
A boolean mask of the cells which were in a dissolved metacell.
If ``inplace`` (default: {inplace}), this is written to the data, and the function returns
``None``. Otherwise this is returned as a pandas data frame (indexed by the observation names).
**Computation Parameters**
1. Mark all cells with non-zero ``deviants`` (default: {deviants}) as "outliers". This can be
the name of a per-observation (cell) annotation, or an explicit boolean mask of cells, or a
or ``None`` if there are no deviant cells to mark.
2. Any metacell which has less cells than the ``min_metacell_cells`` is dissolved.
3. We are trying to create metacells of size ``target_metacell_size``. Compute the sizes of the
resulting metacells by summing the ``cell_sizes`` (default: {cell_sizes}). If it is ``None``,
each has a size of one. These parameters are typically identical to these passed to
:py:func:`metacells.tools.candidates.compute_candidate_metacells`.
4. If ``min_robust_size_factor`` (default: {min_robust_size_factor}) is specified, then any
metacell whose total size is at least ``target_metacell_size * min_robust_size_factor`` is
preserved.
5. If ``min_convincing_size_factor`` (default: {min_convincing_size_factor}) is specified, then any remaining
metacells whose size is at least ``target_metacell_size * min_convincing_size_factor`` are preserved, given they
contain at least one gene whose fold factor (log2((actual + 1) / (expected + 1))) is at least
``min_convincing_gene_fold_factor`` (default: {min_convincing_gene_fold_factor}). If ``abs_folds``, consider the
absolute fold factors. That is, we only preserve these smaller metacells if there is at least one gene whose
expression is significantly different from the mean of the population.
6 . Any remaining metacell is dissolved into "outlier" cells.
"""
dissolved_of_cells = np.zeros(adata.n_obs, dtype="bool")
candidate_of_cells = ut.get_o_numpy(adata,
candidates,
formatter=ut.groups_description)
candidate_of_cells = np.copy(candidate_of_cells)
deviant_of_cells = ut.maybe_o_numpy(adata,
deviants,
formatter=ut.mask_description)
if deviant_of_cells is not None:
deviant_of_cells = deviant_of_cells > 0
cell_sizes = ut.maybe_o_numpy(adata,
cell_sizes,
formatter=ut.sizes_description)
if deviant_of_cells is not None:
candidate_of_cells[deviant_of_cells > 0] = -1
candidate_of_cells = ut.compress_indices(candidate_of_cells)
candidates_count = np.max(candidate_of_cells) + 1
data = ut.get_vo_proper(adata, what, layout="column_major")
fraction_of_genes = ut.fraction_per(data, per="column")
if min_robust_size_factor is None:
min_robust_size = None
else:
min_robust_size = target_metacell_size * min_robust_size_factor
ut.log_calc("min_robust_size", min_robust_size)
if min_convincing_size_factor is None:
min_convincing_size = None
else:
min_convincing_size = target_metacell_size * min_convincing_size_factor
ut.log_calc("min_convincing_size", min_convincing_size)
did_dissolve = False
for candidate_index in range(candidates_count):
candidate_cell_indices = np.where(
candidate_of_cells == candidate_index)[0]
if not _keep_candidate(
adata,
candidate_index,
data=data,
cell_sizes=cell_sizes,
fraction_of_genes=fraction_of_genes,
min_metacell_cells=min_metacell_cells,
min_robust_size=min_robust_size,
min_convincing_size=min_convincing_size,
min_convincing_gene_fold_factor=min_convincing_gene_fold_factor,
abs_folds=abs_folds,
candidates_count=candidates_count,
candidate_cell_indices=candidate_cell_indices,
):
dissolved_of_cells[candidate_cell_indices] = True
candidate_of_cells[candidate_cell_indices] = -1
did_dissolve = True
if did_dissolve:
metacell_of_cells = ut.compress_indices(candidate_of_cells)
else:
metacell_of_cells = candidate_of_cells
if inplace:
ut.set_o_data(adata,
"dissolved",
dissolved_of_cells,
formatter=ut.mask_description)
ut.set_o_data(adata,
"metacell",
metacell_of_cells,
formatter=ut.groups_description)
return None
ut.log_return("dissolved", dissolved_of_cells)
ut.log_return("metacell",
metacell_of_cells,
formatter=ut.groups_description)
obs_frame = ut.to_pandas_frame(index=adata.obs_names)
obs_frame["dissolved"] = dissolved_of_cells
obs_frame["metacell"] = metacell_of_cells
return obs_frame
def _related_genes( # pylint: disable=too-many-statements,too-many-branches
*,
adata_of_all_genes_of_all_cells: AnnData,
what: Union[str, ut.Matrix] = "__x__",
rare_gene_indices_of_modules: List[List[int]],
allowed_genes_mask: ut.NumpyVector,
min_genes_of_modules: int,
min_gene_maximum: int,
min_cells_of_modules: int,
max_cells_of_modules: int,
min_cell_module_total: int,
min_related_gene_fold_factor: float,
max_related_gene_increase_factor: float,
) -> List[List[int]]:
total_all_cells_umis_of_all_genes = ut.get_v_numpy(
adata_of_all_genes_of_all_cells, what, sum=True)
ut.log_calc("genes for modules:")
modules_count = 0
related_gene_indices_of_modules: List[List[int]] = []
rare_gene_indices_of_any: Set[int] = set()
for rare_gene_indices_of_module in rare_gene_indices_of_modules:
if len(rare_gene_indices_of_module) >= min_genes_of_modules:
rare_gene_indices_of_any.update(list(rare_gene_indices_of_module))
for rare_gene_indices_of_module in rare_gene_indices_of_modules:
if len(rare_gene_indices_of_module) < min_genes_of_modules:
continue
module_index = modules_count
modules_count += 1
with ut.log_step("- module", module_index):
ut.log_calc(
"rare_gene_names",
sorted(adata_of_all_genes_of_all_cells.
var_names[rare_gene_indices_of_module]))
adata_of_module_genes_of_all_cells = ut.slice(
adata_of_all_genes_of_all_cells,
name=f".module{module_index}.rare_gene",
vars=rare_gene_indices_of_module,
top_level=False,
)
total_module_genes_umis_of_all_cells = ut.get_o_numpy(
adata_of_module_genes_of_all_cells, what, sum=True)
mask_of_expressed_cells = total_module_genes_umis_of_all_cells > 0
expressed_cells_count = np.sum(mask_of_expressed_cells)
if expressed_cells_count > max_cells_of_modules:
if ut.logging_calc():
ut.log_calc(
"expressed_cells",
ut.mask_description(mask_of_expressed_cells) +
" (too many)")
continue
if expressed_cells_count < min_cells_of_modules:
if ut.logging_calc():
ut.log_calc(
"expressed_cells",
ut.mask_description(mask_of_expressed_cells) +
" (too few)")
continue
ut.log_calc("expressed_cells", mask_of_expressed_cells)
adata_of_all_genes_of_expressed_cells_of_module = ut.slice(
adata_of_all_genes_of_all_cells,
name=f".module{module_index}.rare_cell",
obs=mask_of_expressed_cells,
top_level=False,
)
total_expressed_cells_umis_of_all_genes = ut.get_v_numpy(
adata_of_all_genes_of_expressed_cells_of_module,
what,
sum=True)
data = ut.get_vo_proper(
adata_of_all_genes_of_expressed_cells_of_module,
what,
layout="column_major")
max_expressed_cells_umis_of_all_genes = ut.max_per(data,
per="column")
total_background_cells_umis_of_all_genes = (
total_all_cells_umis_of_all_genes -
total_expressed_cells_umis_of_all_genes)
expressed_cells_fraction_of_all_genes = total_expressed_cells_umis_of_all_genes / sum(
total_expressed_cells_umis_of_all_genes)
background_cells_fraction_of_all_genes = total_background_cells_umis_of_all_genes / sum(
total_background_cells_umis_of_all_genes)
mask_of_related_genes = (
allowed_genes_mask
& (max_expressed_cells_umis_of_all_genes >= min_gene_maximum)
& (expressed_cells_fraction_of_all_genes >=
background_cells_fraction_of_all_genes *
(2**min_related_gene_fold_factor)))
related_gene_indices = np.where(mask_of_related_genes)[0]
assert np.all(mask_of_related_genes[rare_gene_indices_of_module])
base_genes_of_all_cells_adata = ut.slice(
adata_of_all_genes_of_all_cells,
name=f".module{module_index}.base",
vars=rare_gene_indices_of_module)
total_base_genes_of_all_cells = ut.get_o_numpy(
base_genes_of_all_cells_adata, what, sum=True)
mask_of_strong_base_cells = total_base_genes_of_all_cells >= min_cell_module_total
count_of_strong_base_cells = np.sum(mask_of_strong_base_cells)
if ut.logging_calc():
ut.log_calc(
"candidate_gene_names",
sorted(adata_of_all_genes_of_all_cells.
var_names[related_gene_indices]))
ut.log_calc("base_strong_genes", count_of_strong_base_cells)
related_gene_indices_of_module = list(rare_gene_indices_of_module)
for gene_index in related_gene_indices:
if gene_index in rare_gene_indices_of_module:
continue
if gene_index in rare_gene_indices_of_any:
ut.log_calc(
f"- candidate gene {adata_of_all_genes_of_all_cells.var_names[gene_index]} "
f"belongs to another module")
continue
if gene_index not in rare_gene_indices_of_module:
related_gene_of_all_cells_adata = ut.slice(
adata_of_all_genes_of_all_cells,
name=
f".{adata_of_all_genes_of_all_cells.var_names[gene_index]}",
vars=np.array([gene_index]),
)
assert related_gene_of_all_cells_adata.n_vars == 1
total_related_genes_of_all_cells = ut.get_o_numpy(
related_gene_of_all_cells_adata, what, sum=True)
total_related_genes_of_all_cells += total_base_genes_of_all_cells
mask_of_strong_related_cells = total_related_genes_of_all_cells >= min_cell_module_total
count_of_strong_related_cells = np.sum(
mask_of_strong_related_cells)
ut.log_calc(
f"- candidate gene {adata_of_all_genes_of_all_cells.var_names[gene_index]} "
f"strong cells: {count_of_strong_related_cells} "
f"factor: {count_of_strong_related_cells / count_of_strong_base_cells}"
)
if count_of_strong_related_cells > max_related_gene_increase_factor * count_of_strong_base_cells:
continue
related_gene_indices_of_module.append(gene_index)
related_gene_indices_of_modules.append(
related_gene_indices_of_module) #
if ut.logging_calc():
ut.log_calc("related genes for modules:")
for module_index, related_gene_indices_of_module in enumerate(
related_gene_indices_of_modules):
ut.log_calc(
f"- module {module_index} related_gene_names",
sorted(adata_of_all_genes_of_all_cells.
var_names[related_gene_indices_of_module]),
)
return related_gene_indices_of_modules
def find_systematic_genes(
what: Union[str, ut.Matrix] = "__x__",
*,
adata: AnnData,
qdata: AnnData,
atlas_total_umis: Optional[ut.Vector] = None,
query_total_umis: Optional[ut.Vector] = None,
low_gene_quantile: float = pr.systematic_low_gene_quantile,
high_gene_quantile: float = pr.systematic_high_gene_quantile,
to_property_name: str = "systematic_gene",
) -> None:
"""
Find genes that
**Input**
Annotated query ``qdata`` and atlas ``adata``, where the observations are cells and the variables are genes, where
``what`` is a per-variable-per-observation matrix or the name of a per-variable-per-observation annotation
containing such a matrix.
**Returns**
A matrix whose rows are query metacells and columns are atlas metacells, where each entry is the weight of the atlas
metacell in the projection of the query metacells. The sum of weights in each row (that is, for a single query
metacell) is 1. The weighted sum of the atlas metacells using these weights is the "projected" image of the query
metacell onto the atlas.
In addition, sets the following annotations in ``qdata``:
Variable (Gene) Annotations
``systematic_gene`` (or ``to_property_name``)
A boolean mask indicating whether the gene is systematically higher or lower in the query compared to the
atlas.
**Computation Parameters**
1. Compute the fraction of each gene out of the total UMIs in both the atlas and the query. If ``atlas_total_umis``
and/or ``query_total_umis`` are given, use them as the basis instead of the sum of the UMIs.
2. Compute for each gene its ``low_gene_quantile`` (default: {low_gene_quantile}) fraction in the query, and its
``high_gene_quantile`` (default: {high_gene_quantile}) fraction in the atlas.
3. Compute for each gene its standard deviation in the atlas.
4. Mark as systematic the genes for which the low quantile value in the query is at least the atlas high quantile
value.
5. Mark as systematic the genes for which the low quantile value in the atlas is at least the query high quantile
value.
"""
assert 0 <= low_gene_quantile <= 1
assert 0 <= high_gene_quantile <= 1
assert np.all(adata.var_names == qdata.var_names)
query_umis = ut.get_vo_proper(qdata, what, layout="row_major")
atlas_umis = ut.get_vo_proper(adata, what, layout="row_major")
atlas_fractions = ut.to_numpy_matrix(ut.fraction_by(atlas_umis, by="row", sums=atlas_total_umis))
query_fractions = ut.to_numpy_matrix(ut.fraction_by(query_umis, by="row", sums=query_total_umis))
query_fractions = ut.to_layout(query_fractions, layout="column_major")
atlas_fractions = ut.to_layout(atlas_fractions, layout="column_major")
query_low_gene_values = ut.quantile_per(query_fractions, low_gene_quantile, per="column")
atlas_low_gene_values = ut.quantile_per(atlas_fractions, low_gene_quantile, per="column")
query_high_gene_values = ut.quantile_per(query_fractions, high_gene_quantile, per="column")
atlas_high_gene_values = ut.quantile_per(atlas_fractions, high_gene_quantile, per="column")
query_above_atlas = query_low_gene_values > atlas_high_gene_values
atlas_above_query = atlas_low_gene_values >= query_high_gene_values
systematic = query_above_atlas | atlas_above_query
ut.set_v_data(qdata, to_property_name, systematic)
def project_query_onto_atlas(
what: Union[str, ut.Matrix] = "__x__",
*,
adata: AnnData,
qdata: AnnData,
atlas_total_umis: Optional[ut.Vector] = None,
query_total_umis: Optional[ut.Vector] = None,
project_log_data: bool = pr.project_log_data,
fold_normalization: float = pr.project_fold_normalization,
min_significant_gene_value: float = pr.project_min_significant_gene_value,
max_consistency_fold_factor: float = pr.project_max_consistency_fold_factor,
candidates_count: int = pr.project_candidates_count,
min_usage_weight: float = pr.project_min_usage_weight,
reproducible: bool,
) -> ut.CompressedMatrix:
"""
Project query metacells onto atlas metacells.
**Input**
Annotated query ``qdata`` and atlas ``adata``, where the observations are cells and the variables are genes, where
``what`` is a per-variable-per-observation matrix or the name of a per-variable-per-observation annotation
containing such a matrix.
Typically this data excludes any genes having a systematic difference between the query and the atlas, e.g. genes
detected by by :py:func:`metacells.tools.project.find_systematic_genes`.
**Returns**
A matrix whose rows are query metacells and columns are atlas metacells, where each entry is the weight of the atlas
metacell in the projection of the query metacells. The sum of weights in each row (that is, for a single query
metacell) is 1. The weighted sum of the atlas metacells using these weights is the "projected" image of the query
metacell onto the atlas.
In addition, sets the following annotations in ``qdata``:
Observation (Cell) Annotations
``similar``
A boolean mask indicating whether the query metacell is similar to its projection onto the atlas. If
``False`` the metacells is said to be "dissimilar", which may indicate the query contains cell states that
do not appear in the atlas.
**Computation Parameters**
0. All fold computations (log2 of the ratio between gene expressions as a fraction of the total UMIs) use the
``fold_normalization`` (default: {fold_normalization}). Fractions are computed based on the total UMIs, unless
``atlas_total_umis`` and/or ``query_total_umis`` are specified.
For each query metacell:
1. Correlate the metacell with all the atlas metacells, and pick the highest-correlated one as the "anchor".
If ``reproducible``, a slower (still parallel) but reproducible algorithm will be used.
2. Consider as candidates only atlas metacells whose maximal gene fold factor compared to the anchor is at most
``max_consistency_fold_factor`` (default: {max_consistency_fold_factor}). Ignore the fold factors of genes whose
sum of UMIs in the anchor and the candidate metacells is less than ``min_significant_gene_value`` (default:
{min_significant_gene_value}).
3. Select the ``candidates_count`` (default: {candidates_count}) candidate metacells with the highest correlation
with the query metacell.
4. Compute the non-negative weights (with a sum of 1) of the selected candidates that give the best projection of
the query metacells onto the atlas. Since the algorithm for computing these weights rarely produces an exact 0
weight, reduce all weights less than the ``min_usage_weight`` (default: {min_usage_weight}) to zero. If
``project_log_data`` (default: {project_log_data}), compute the match on the log of the data instead of the
actual data.
"""
assert fold_normalization > 0
assert candidates_count > 0
assert min_usage_weight >= 0
assert max_consistency_fold_factor >= 0
assert np.all(adata.var_names == qdata.var_names)
atlas_umis = ut.get_vo_proper(adata, what, layout="row_major")
query_umis = ut.get_vo_proper(qdata, what, layout="row_major")
if atlas_total_umis is None:
atlas_total_umis = ut.sum_per(atlas_umis, per="row")
atlas_total_umis = ut.to_numpy_vector(atlas_total_umis)
if query_total_umis is None:
query_total_umis = ut.sum_per(query_umis, per="row")
query_total_umis = ut.to_numpy_vector(query_total_umis)
atlas_fractions = ut.to_numpy_matrix(ut.fraction_by(atlas_umis, by="row", sums=atlas_total_umis))
query_fractions = ut.to_numpy_matrix(ut.fraction_by(query_umis, by="row", sums=query_total_umis))
atlas_fractions += fold_normalization
query_fractions += fold_normalization
atlas_log_fractions = np.log2(atlas_fractions)
query_log_fractions = np.log2(query_fractions)
atlas_fractions -= fold_normalization
query_fractions -= fold_normalization
if project_log_data:
atlas_project_data = atlas_log_fractions
query_project_data = query_log_fractions
else:
atlas_project_data = atlas_fractions
query_project_data = query_fractions
query_atlas_corr = ut.cross_corrcoef_rows(query_project_data, atlas_project_data, reproducible=reproducible)
@ut.timed_call("project_single_metacell")
def _project_single(query_metacell_index: int) -> Tuple[ut.NumpyVector, ut.NumpyVector]:
return _project_single_metacell(
atlas_umis=atlas_umis,
query_atlas_corr=query_atlas_corr,
atlas_project_data=atlas_project_data,
query_project_data=query_project_data,
atlas_log_fractions=atlas_log_fractions,
candidates_count=candidates_count,
min_significant_gene_value=min_significant_gene_value,
min_usage_weight=min_usage_weight,
max_consistency_fold_factor=max_consistency_fold_factor,
query_metacell_index=query_metacell_index,
)
results = list(ut.parallel_map(_project_single, qdata.n_obs))
indices = np.concatenate([result[0] for result in results], dtype="int32")
data = np.concatenate([result[1] for result in results], dtype="float32")
atlas_used_sizes = [len(result[0]) for result in results]
atlas_used_sizes.insert(0, 0)
indptr = np.cumsum(np.array(atlas_used_sizes))
return sp.csr_matrix((data, indices, indptr), shape=(qdata.n_obs, adata.n_obs))
def compute_inner_fold_factors(
what: Union[str, ut.Matrix] = "__x__",
*,
adata: AnnData,
gdata: AnnData,
group: Union[str, ut.Vector] = "metacell",
min_gene_inner_fold_factor: float = pr.min_gene_inner_fold_factor,
min_entry_inner_fold_factor: float = pr.min_entry_inner_fold_factor,
inner_abs_folds: float = pr.inner_abs_folds,
) -> None:
"""
Compute the inner fold factors of genes within in each metacell.
This computes, for each cell of the metacell, the same fold factors that are used to detect deviant cells (see
:py:func:`metacells.tools.deviants.find_deviant_cells`), and keeps the maximal fold for each gene in the metacell.
The result per-metacell-per-gene matrix is then made sparse by discarding too-low values (setting them to zero).
Ideally, this matrix should be "very" sparse. If it contains "too many" non-zero values, this indicates the
metacells contains "too much" variability. This may be due to actual biology (e.g. immune cells or olfactory nerves
which are all similar except for each one expressing one different gene), due to batch effects (similar cells in
distinct batches differing in some genes due to technical issues), due to low data quality (the overall noise level
is so high that this is simply the best the algorithm can do), or worse - a combination of the above.
**Input**
Annotated ``adata``, where the observations are cells and the variables are genes, where ``what`` is a
per-variable-per-observation matrix or the name of a per-variable-per-observation annotation containing such a
matrix.
In addition, ``gdata`` is assumed to have one observation for each group, and use the same
genes as ``adata``.
**Returns**
Sets the following in ``gdata``:
Per-Variable Per-Observation (Gene-Cell) Annotations
``inner_fold``
For each gene and group, the maximal fold factor of this gene in any cell contained in the group (unless the
value is too low to be of interest, in which case it will be zero).
**Computation Parameters**
1. For each group (metacell), for each gene, compute the gene's maximal (in all the cells of the group) fold factor
log2((actual UMIs + 1) / (expected UMIs + 1)), similarly to
:py:func:`metacells.tools.deviants.find_deviant_cells`.
2. If the maximal fold factor for a gene (across all metacells) is below ``min_gene_inner_fold_factor`` (default:
{min_gene_inner_fold_factor}), then set all the gene's fold factors to zero (too low to be of interest). If
``inner_abs_folds`` (default: {inner_abs_folds}), consider the absolute fold factors.
3. Otherwise, for any metacell whose fold factor for the gene is less than ``min_entry_inner_fold_factor`` (default:
{min_entry_inner_fold_factor}), set the fold factor to zero (too low to be of interest).
"""
assert 0 <= min_entry_inner_fold_factor <= min_gene_inner_fold_factor
cells_data = ut.get_vo_proper(adata, what, layout="row_major")
metacells_data = ut.get_vo_proper(gdata, what, layout="row_major")
group_of_cells = ut.get_o_numpy(adata,
group,
formatter=ut.groups_description)
total_umis_per_cell = ut.sum_per(cells_data, per="row")
total_umis_per_metacell = ut.sum_per(metacells_data, per="row")
@ut.timed_call("compute_metacell_inner_folds")
def _compute_single_metacell_inner_folds(
metacell_index: int) -> ut.NumpyVector:
return _compute_metacell_inner_folds(
metacell_index=metacell_index,
cells_data=cells_data,
metacells_data=metacells_data,
group_of_cells=group_of_cells,
total_umis_per_cell=total_umis_per_cell,
total_umis_per_metacell=total_umis_per_metacell,
)
results = list(
ut.parallel_map(_compute_single_metacell_inner_folds, gdata.n_obs))
dense_inner_folds_by_row = np.array(results)
dense_inner_folds_by_column = ut.to_layout(dense_inner_folds_by_row,
"column_major")
if inner_abs_folds:
comparable_dense_inner_folds_by_column = np.abs(
dense_inner_folds_by_column)
else:
comparable_dense_inner_folds_by_column = dense_inner_folds_by_column
max_fold_per_gene = ut.max_per(comparable_dense_inner_folds_by_column,
per="column")
significant_genes_mask = max_fold_per_gene >= min_gene_inner_fold_factor
ut.log_calc("significant_genes_mask", significant_genes_mask)
dense_inner_folds_by_column[:, ~significant_genes_mask] = 0
dense_inner_folds_by_column[comparable_dense_inner_folds_by_column <
min_entry_inner_fold_factor] = 0
dense_inner_folds_by_row = ut.to_layout(dense_inner_folds_by_column,
layout="row_major")
sparse_inner_folds = sparse.csr_matrix(dense_inner_folds_by_row)
ut.set_vo_data(gdata, "inner_fold", sparse_inner_folds)
def group_obs_data(
adata: AnnData,
what: Union[str, ut.Matrix] = "__x__",
*,
groups: Union[str, ut.Vector],
name: Optional[str] = None,
) -> Optional[AnnData]:
"""
Compute new data which has the ``what`` (default: {what}) sum of the observations (cells) for
each group.
For example, having computed a metacell index for each cell, compute the per-metacell data
for further analysis.
If ``groups`` is a string, it is expected to be the name of a per-observation vector annotation.
Otherwise it should be a vector. The group indices should be integers, where negative values
indicate "no group" and non-negative values indicate the index of the group to which each
observation (cell) belongs to.
**Input**
Annotated ``adata``, where the observations are cells and the variables are genes, where
``what`` is a per-variable-per-observation matrix or the name of a per-variable-per-observation
annotation containing such a matrix.
**Returns**
An annotated data where each observation is the sum of the group of original observations
(cells). Observations with a negative group index are discarded. If all observations are
discarded, return ``None``.
The new data will contain only:
* An ``X`` member holding the summed-per-group data.
* A new ``grouped`` per-observation data which counts, for each group, the number
of grouped observations summed into it.
If ``name`` is not specified, the data will be unnamed. Otherwise, if it starts with a ``.``, it
will be appended to the current name (if any). Otherwise, ``name`` is the new name.
"""
group_of_cells = ut.get_o_numpy(adata,
groups,
formatter=ut.groups_description)
data = ut.get_vo_proper(adata, what, layout="row_major")
results = ut.sum_groups(data, group_of_cells, per="row")
if results is None:
return None
summed_data, cell_counts = results
gdata = AnnData(summed_data)
gdata.var_names = adata.var_names
ut.set_name(gdata, ut.get_name(adata))
ut.set_name(gdata, name)
ut.set_o_data(gdata,
"grouped",
cell_counts,
formatter=ut.sizes_description)
return gdata
def collect_metacells(
adata: AnnData,
what: Union[str, ut.Matrix] = "__x__",
*,
max_cell_size: Optional[float] = pr.max_cell_size,
max_cell_size_factor: Optional[float] = pr.max_cell_size_factor,
cell_sizes: Optional[Union[str, ut.Vector]] = pr.cell_sizes,
name: str = "metacells",
top_level: bool = True,
) -> AnnData:
"""
Collect computed metacells ``what`` (default: {what}) data.
**Input**
Annotated (presumably "clean") ``adata``, where the observations are cells and the variables are
genes, and where ``what`` is a per-variable-per-observation matrix or the name of a
per-variable-per-observation annotation containing such a matrix.
**Returns**
Annotated metacell data containing for each observation the sum of the data (by of the cells for
each metacell, which contains the following annotations:
Variable (Gene) Annotations
``excluded_gene``
A mask of the genes which were excluded by name.
``clean_gene``
A boolean mask of the clean genes.
``forbidden_gene``
A boolean mask of genes which are forbidden from being chosen as "feature" genes based
on their name. This is ``False`` for non-"clean" genes.
If directly computing metecalls:
``feature``
A boolean mask of the "feature" genes. This is ``False`` for non-"clean" genes.
If using divide-and-conquer:
``pre_feature``, ``feature``
The number of times the gene was used as a feature when computing the preliminary and
final metacells. This is zero for non-"clean" genes.
Observations (Cell) Annotations
``grouped``
The number of ("clean") cells grouped into each metacell.
``pile``
The index of the pile used to compute the metacell each cell was assigned to to. This is
``-1`` for non-"clean" cells.
``candidate``
The index of the candidate metacell each cell was assigned to to. This is ``-1`` for
non-"clean" cells.
Also sets all relevant annotations in the full data based on their value in the clean data, with
appropriate defaults for non-"clean" data.
**Computation Parameters**
1. Compute the cell's scale factors by invoking :py:func:`compute_effective_cell_sizes` using the
``max_cell_size`` (default: {max_cell_size}), ``max_cell_size_factor`` (default:
{max_cell_size_factor}) and ``cell_sizes`` (default: {cell_sizes}).
2. Scale the cell's data using these factors, if needed.
3. Invoke :py:func:`metacells.tools.group.group_obs_data` to sum the cells into
metacells.
4. Pass all relevant per-gene and per-cell annotations to the result.
"""
_cell_sizes, _max_cell_size, cell_scale_factors = compute_effective_cell_sizes(
adata, max_cell_size=max_cell_size, max_cell_size_factor=max_cell_size_factor, cell_sizes=cell_sizes
)
if cell_scale_factors is not None:
data = ut.get_vo_proper(adata, what, layout="row_major")
what = ut.scale_by(data, cell_scale_factors, by="row")
mdata = tl.group_obs_data(adata, what, groups="metacell", name=name)
assert mdata is not None
if top_level:
ut.top_level(mdata)
for annotation_name in ("excluded_gene", "clean_gene", "forbidden_gene", "pre_feature_gene", "feature_gene"):
if not ut.has_data(adata, annotation_name):
continue
value_per_gene = ut.get_v_numpy(adata, annotation_name, formatter=ut.mask_description)
ut.set_v_data(mdata, annotation_name, value_per_gene, formatter=ut.mask_description)
for annotation_name in ("pile", "candidate"):
if ut.has_data(adata, annotation_name):
tl.group_obs_annotation(
adata, mdata, groups="metacell", formatter=ut.groups_description, name=annotation_name, method="unique"
)
return mdata
def _compute_elements_similarity( # pylint: disable=too-many-branches
adata: AnnData,
elements: str,
per: str,
what: Union[str, ut.Matrix],
*,
method: str,
reproducible: bool,
logistics_location: float,
logistics_slope: float,
top: Optional[int],
bottom: Optional[int],
inplace: bool,
) -> Optional[ut.PandasFrame]:
assert elements in ("obs", "var")
assert method in ("pearson", "repeated_pearson", "logistics",
"logistics_pearson")
data = ut.get_vo_proper(adata, what, layout=f"{per}_major")
dense = ut.to_numpy_matrix(data)
similarity: ut.ProperMatrix
if method.startswith("logistics"):
similarity = ut.logistics(dense,
location=logistics_location,
slope=logistics_slope,
per=per)
similarity *= -1
similarity += 1
else:
similarity = ut.corrcoef(dense, per=per, reproducible=reproducible)
if method.endswith("_pearson"):
similarity = ut.corrcoef(similarity,
per=None,
reproducible=reproducible)
if top is not None:
top_similarity = ut.top_per(similarity, top, per="row")
if bottom is not None:
similarity *= -1
bottom_similarity = ut.top_per(similarity, bottom, per="row")
bottom_similarity *= -1 # type: ignore
if top is not None:
if bottom is not None:
similarity = top_similarity + bottom_similarity # type: ignore
else:
similarity = top_similarity
else:
if bottom is not None:
similarity = bottom_similarity
if inplace:
to = elements + "_similarity"
if elements == "obs":
ut.set_oo_data(adata, to, similarity)
else:
ut.set_vv_data(adata, to, similarity)
return None
if elements == "obs":
names = adata.obs_names
else:
names = adata.var_names
return ut.to_pandas_frame(similarity, index=names, columns=names)
def compute_knn_by_features(
adata: AnnData,
what: Union[str, ut.Matrix] = "__x__",
*,
max_top_feature_genes: int = pr.max_top_feature_genes,
similarity_value_normalization: float = pr.
umap_similarity_value_normalization,
similarity_log_data: bool = pr.umap_similarity_log_data,
similarity_method: str = pr.umap_similarity_method,
logistics_location: float = pr.logistics_location,
logistics_slope: float = pr.logistics_slope,
k: int,
balanced_ranks_factor: float = pr.knn_balanced_ranks_factor,
incoming_degree_factor: float = pr.knn_incoming_degree_factor,
outgoing_degree_factor: float = pr.knn_outgoing_degree_factor,
reproducible: bool = pr.reproducible,
) -> ut.PandasFrame:
"""
Compute KNN graph between metacells based on feature genes.
If ``reproducible`` (default: {reproducible}) is ``True``, a slower (still parallel) but
reproducible algorithm will be used to compute pearson correlations.
**Input**
Annotated ``adata`` where each observation is a metacells and the variables are genes,
are genes, where ``what`` is a per-variable-per-observation matrix or the name of a
per-variable-per-observation annotation containing such a matrix.
**Returns**
Sets the following in ``adata``:
Observations-Pair (Metacells) Annotations
``obs_outgoing_weights``
A sparse square matrix where each non-zero entry is the weight of an edge between a pair
of cells or genes, where the sum of the weights of the outgoing edges for each element
is 1 (there is always at least one such edge).
Also return a pandas data frame of the similarities between the observations (metacells).
**Computation Parameters**
1. Invoke :py:func:`metacells.tools.high.find_top_feature_genes` using ``max_top_feature_genes``
(default: {max_top_feature_genes}) to pick the feature genes to use to compute similarities
between the metacells.
2. Compute the fractions of each gene in each cell, and add the
``similarity_value_normalization`` (default: {similarity_value_normalization}) to
it.
3. If ``similarity_log_data`` (default: {similarity_log_data}), invoke the
:py:func:`metacells.utilities.computation.log_data` function to compute the log (base 2) of
the data.
4. Invoke :py:func:`metacells.tools.similarity.compute_obs_obs_similarity` using
``similarity_method`` (default: {similarity_method}), ``logistics_location`` (default:
{logistics_slope}) and ``logistics_slope`` (default: {logistics_slope}) and convert this
to distances.
5. Invoke :py:func:`metacells.tools.knn_graph.compute_obs_obs_knn_graph` using the distances,
``k`` (no default!), ``balanced_ranks_factor`` (default: {balanced_ranks_factor}),
``incoming_degree_factor`` (default: {incoming_degree_factor}), ``outgoing_degree_factor``
(default: {outgoing_degree_factor}) to compute a "skeleton" graph to overlay on top of the
UMAP graph.
"""
tl.find_top_feature_genes(adata, max_genes=max_top_feature_genes)
all_data = ut.get_vo_proper(adata, what, layout="row_major")
all_fractions = ut.fraction_by(all_data, by="row")
top_feature_genes_mask = ut.get_v_numpy(adata, "top_feature_gene")
top_feature_genes_fractions = all_fractions[:, top_feature_genes_mask]
top_feature_genes_fractions = ut.to_layout(top_feature_genes_fractions,
layout="row_major")
top_feature_genes_fractions = ut.to_numpy_matrix(
top_feature_genes_fractions)
top_feature_genes_fractions += similarity_value_normalization
if similarity_log_data:
top_feature_genes_fractions = ut.log_data(top_feature_genes_fractions,
base=2)
tdata = ut.slice(adata, vars=top_feature_genes_mask)
similarities = tl.compute_obs_obs_similarity(
tdata,
top_feature_genes_fractions,
method=similarity_method,
reproducible=reproducible,
logistics_location=logistics_location,
logistics_slope=logistics_slope,
inplace=False,
)
assert similarities is not None
tl.compute_obs_obs_knn_graph(
adata,
similarities,
k=k,
balanced_ranks_factor=balanced_ranks_factor,
incoming_degree_factor=incoming_degree_factor,
outgoing_degree_factor=outgoing_degree_factor,
)
return similarities
def compute_deviant_fold_factors(
what: Union[str, ut.Matrix] = "__x__",
*,
adata: AnnData,
gdata: AnnData,
group: Union[str, ut.Vector] = "metacell",
similar: Union[str, ut.Vector] = "similar",
significant_gene_fold_factor: float = pr.significant_gene_fold_factor,
) -> None:
"""
Given an assignment of observations (cells) to groups (metacells) or, if an outlier, to the most
similar groups, compute for each observation and gene the fold factor relative to its group
for the purpose of detecting deviant cells.
Ideally, all grouped cells would have no genes with high enough fold factors to be considered deviants, and all
outlier cells would. In practice grouped cells might have a (few) such genes to the restriction on the fraction
of deviants.
It is important not to read too much into the results for a single cell, but looking at which genes appear for cell
populations (e.g., cells with specific metadata such as batch identification) might be instructive.
**Input**
Annotated ``adata``, where the observations are cells and the variables are genes, where ``what`` is a
per-variable-per-observation matrix or the name of a per-variable-per-observation annotation containing such a
matrix.
In addition, ``gdata`` is assumed to have one observation for each group, and use the same genes as ``adata``.
**Returns**
Sets the following in ``adata``:
Per-Variable Per-Observation (Gene-Cell) Annotations
``deviant_fold``
The fold factor between the cell's UMIs and the expected number of UMIs for the purpose of computing
deviant cells.
**Computation Parameters**
1. For each cell, compute the expected UMIs for each gene given the fraction of the gene in the metacells associated
with the cell (the one it is belongs to, or the most similar one for outliers). If this is less than
``significant_gene_fold_factor`` (default: {significant_gene_fold_factor}), set it to zero so the result will be
sparse.
"""
cells_data = ut.get_vo_proper(adata, what, layout="row_major")
metacells_data = ut.get_vo_proper(gdata, what, layout="row_major")
total_umis_per_cell = ut.sum_per(cells_data, per="row")
total_umis_per_metacell = ut.sum_per(metacells_data, per="row")
group_of_cells = ut.get_o_numpy(adata,
group,
formatter=ut.groups_description)
similar_of_cells = ut.get_o_numpy(adata,
similar,
formatter=ut.groups_description)
@ut.timed_call("compute_cell_deviant_certificates")
def _compute_cell_deviant_certificates(
cell_index: int) -> Tuple[ut.NumpyVector, ut.NumpyVector]:
return _compute_cell_certificates(
cell_index=cell_index,
cells_data=cells_data,
metacells_data=metacells_data,
group_of_cells=group_of_cells,
similar_of_cells=similar_of_cells,
total_umis_per_cell=total_umis_per_cell,
total_umis_per_metacell=total_umis_per_metacell,
significant_gene_fold_factor=significant_gene_fold_factor,
)
results = list(
ut.parallel_map(_compute_cell_deviant_certificates, adata.n_obs))
cell_indices = np.concatenate([
np.full(len(result[0]), cell_index, dtype="int32")
for cell_index, result in enumerate(results)
])
gene_indices = np.concatenate([result[0] for result in results])
fold_factors = np.concatenate([result[1] for result in results])
deviant_folds = sparse.csr_matrix(
(fold_factors, (cell_indices, gene_indices)), shape=adata.shape)
ut.set_vo_data(adata, "deviant_folds", deviant_folds)
def compute_outliers_matches(
what: Union[str, ut.Matrix] = "__x__",
*,
adata: AnnData,
gdata: AnnData,
group: Union[str, ut.Vector] = "metacell",
similar: str = "similar",
value_normalization: float = pr.outliers_value_normalization,
reproducible: bool,
) -> None:
"""
Given an assignment of observations (cells) to groups (metacells), compute for each outlier the "most similar"
group.
**Input**
Annotated ``adata``, where the observations are cells and the variables are genes, where ``what`` is a
per-variable-per-observation matrix or the name of a per-variable-per-observation annotation containing such a
matrix.
In addition, ``gdata`` is assumed to have one observation for each group, and use the same genes as ``adata``.
**Returns**
Sets the following in ``adata``:
Per-Observation (Cell) Annotations
``similar`` (default: {similar})
For each observation (cell), the index of the "most similar" group.
**Computation Parameters**
1. Compute the log2 of the fraction of each gene in each of the outlier cells and the group metacells using
the ``value_normalization`` (default: {value_normalization}).
2. Cross-correlate each of the outlier cells with each of the group metacells, in a ``reproducible`` manner.
"""
group_of_cells = ut.get_o_numpy(adata, group)
outliers_mask = group_of_cells < 0
odata = ut.slice(adata, obs=outliers_mask)
outliers_data = ut.get_vo_proper(odata, what, layout="row_major")
groups_data = ut.get_vo_proper(gdata, what, layout="row_major")
outliers_fractions = ut.fraction_by(outliers_data, by="row")
groups_fractions = ut.fraction_by(groups_data, by="row")
outliers_fractions = ut.to_numpy_matrix(outliers_fractions)
groups_fractions = ut.to_numpy_matrix(groups_fractions)
outliers_fractions += value_normalization
groups_fractions += value_normalization
outliers_log_fractions = np.log2(outliers_fractions,
out=outliers_fractions)
groups_log_fractions = np.log2(groups_fractions, out=groups_fractions)
outliers_groups_correlation = ut.cross_corrcoef_rows(
outliers_log_fractions,
groups_log_fractions,
reproducible=reproducible)
outliers_similar_group_indices = np.argmax(outliers_groups_correlation,
axis=1)
assert len(outliers_similar_group_indices) == odata.n_obs
cells_similar_group_indices = np.full(adata.n_obs, -1, dtype="int32")
cells_similar_group_indices[outliers_mask] = outliers_similar_group_indices
ut.set_o_data(adata, similar, cells_similar_group_indices)
def compute_subset_distinct_genes(
adata: AnnData,
what: Union[str, ut.Matrix] = "__x__",
*,
prefix: Optional[str] = None,
scale: Optional[Union[bool, str, ut.NumpyVector]],
subset: Union[str, ut.NumpyVector],
normalization: float,
) -> Optional[Tuple[ut.PandasSeries, ut.PandasSeries]]:
"""
Given a subset of the observations (cells), compute for each gene how distinct its ``what``
(default: {what}) value is in the subset compared to the overall population.
This is the area-under-curve of the receiver operating characteristic (AUROC) for the gene, that
is, the probability that a randomly selected observation (cell) in the subset will have a higher
value than a randomly selected observation (cell) outside the subset.
**Input**
Annotated ``adata``, where the observations are cells and the variables are genes, where
``what`` is a per-variable-per-observation matrix or the name of a per-variable-per-observation
annotation containing such a matrix.
**Returns**
Variable (Gene) Annotations
``<prefix>_fold``
Store the ratio of the expression of the gene in the subset as opposed to the rest of
the population.
``<prefix>_auroc``
Store the distinctiveness of the gene in the subset as opposed to the rest of the
population.
If ``prefix`` (default: {prefix}), is specified, this is written to the data. Otherwise this is
returned as two pandas series (indexed by the gene names).
**Computation Parameters**
1. Use the ``subset`` to assign a boolean label to each observation (cell). The ``subset`` can
be a vector of integer observation names, or a boolean mask, or the string name of a
per-observation annotation containing the boolean mask.
2. If ``scale`` is ``False``, use the data as-is. If it is ``True``, divide the data by the
sum of each observation (cell). If it is a string, it should be the name of a per-observation
annotation to use. Otherwise, it should be a vector of the scale factor for each observation
(cell).
3. Compute the fold ratios using the ``normalization`` (no default!) and the AUROC for each gene,
for the scaled data based on this mask.
"""
if isinstance(subset, str):
subset = ut.get_o_numpy(adata, subset)
if subset.dtype != "bool":
mask: ut.NumpyVector = np.full(adata.n_obs, False)
mask[subset] = True
subset = mask
scale_of_cells: Optional[ut.NumpyVector] = None
if not isinstance(scale, bool):
scale_of_cells = ut.maybe_o_numpy(adata,
scale,
formatter=ut.sizes_description)
elif scale:
scale_of_cells = ut.get_o_numpy(adata, what, sum=True)
else:
scale_of_cells = None
matrix = ut.get_vo_proper(adata, what, layout="column_major").transpose()
fold_of_genes, auroc_of_genes = ut.matrix_rows_folds_and_aurocs(
matrix,
columns_subset=subset,
columns_scale=scale_of_cells,
normalization=normalization)
if prefix is not None:
ut.set_v_data(adata, f"{prefix}_auroc", auroc_of_genes)
ut.set_v_data(adata, f"{prefix}_fold", fold_of_genes)
return None
return (
ut.to_pandas_series(fold_of_genes, index=adata.var_names),
ut.to_pandas_series(auroc_of_genes, index=adata.var_names),
)
def compute_distinct_folds(
adata: AnnData,
what: Union[str, ut.Matrix] = "__x__",
*,
normalization: float = 0,
inplace: bool = True,
) -> Optional[ut.PandasFrame]:
"""
Compute for each observation (cell) and each variable (gene) how much is the ``what`` (default:
{what}) value different from the overall population.
**Input**
Annotated ``adata``, where the observations are cells and the variables are genes, where
``what`` is a per-variable-per-observation matrix or the name of a per-variable-per-observation
annotation containing such a matrix.
**Returns**
Per-Observation-Per-Variable (Cell-Gene) Annotations:
``distinct_ratio``
For each gene in each cell, the log based 2 of the ratio between the fraction of the
gene in the cell and the fraction of the gene in the overall population (sum of cells).
If ``inplace`` (default: {inplace}), this is written to the data, and the function returns
``None``. Otherwise this is returned as a pandas frame (indexed by the observation and
distinct gene rank).
**Computation Parameters**
1. Compute, for each gene, the fraction of the gene's values out of the total sum of the values
(that is, the mean fraction of the gene's expression in the population).
2. Compute, for each cell, for each gene, the fraction of the gene's value out of the sum
of the values in the cell (that is, the fraction of the gene's expression in the cell).
3. Divide the two to the distinct ratio (that is, how much the gene's expression in the cell is
different from the overall population), first adding the ``normalization`` (default:
{normalization}) to both.
4. Compute the log (base 2) of the result and use it as the fold factor.
"""
columns_data = ut.get_vo_proper(adata, what, layout="column_major")
fractions_of_genes_in_data = ut.fraction_per(columns_data, per="column")
fractions_of_genes_in_data += normalization
total_umis_of_cells = ut.get_o_numpy(adata, what, sum=True)
total_umis_of_cells[total_umis_of_cells == 0] = 1
rows_data = ut.get_vo_proper(adata, what, layout="row_major")
fraction_of_genes_in_cells = ut.to_numpy_matrix(
rows_data) / total_umis_of_cells[:, np.newaxis]
fraction_of_genes_in_cells += normalization
zeros_mask = fractions_of_genes_in_data <= 0
fractions_of_genes_in_data[zeros_mask] = -1
fraction_of_genes_in_cells[:, zeros_mask] = -1
ratio_of_genes_in_cells = fraction_of_genes_in_cells
ratio_of_genes_in_cells /= fractions_of_genes_in_data
assert np.min(np.min(ratio_of_genes_in_cells)) > 0
fold_of_genes_in_cells = np.log2(ratio_of_genes_in_cells,
out=ratio_of_genes_in_cells)
if inplace:
ut.set_vo_data(adata, "distinct_fold", fold_of_genes_in_cells)
return None
return ut.to_pandas_frame(fold_of_genes_in_cells,
index=adata.obs_names,
columns=adata.var_names)
def find_properly_sampled_cells(
adata: AnnData,
what: Union[str, ut.Matrix] = "__x__",
*,
min_cell_total: Optional[int],
max_cell_total: Optional[int],
excluded_adata: Optional[AnnData] = None,
max_excluded_genes_fraction: Optional[float],
inplace: bool = True,
) -> Optional[ut.PandasSeries]:
"""
Detect cells with a "proper" amount of ``what`` (default: {what}) data.
Due to both technical effects and natural variance between cells, the total number of UMIs
varies from cell to cell. We often would like to work on cells that contain a sufficient number
of UMIs for meaningful analysis; we sometimes also wish to exclude cells which have "too many"
UMIs.
**Input**
Annotated ``adata``, where the observations are cells and the variables are genes, where
``what`` is a per-variable-per-observation matrix or the name of a per-variable-per-observation
annotation containing such a matrix.
**Returns**
Observation (Cell) Annotations
``properly_sampled_cell``
A boolean mask indicating whether each cell has a "proper" amount of UMIs.
If ``inplace`` (default: {inplace}), this is written to the data, and the function returns
``None``. Otherwise this is returned as a pandas series (indexed by the observation names).
**Computation Parameters**
1. Exclude all cells whose total data is less than the ``min_cell_total`` (no default), unless
it is ``None``.
2. Exclude all cells whose total data is more than the ``max_cell_total`` (no default), unless
it is ``None``.
3. If ``max_excluded_genes_fraction`` (no default) is not ``None``, then ``excluded_adata`` must
not be ``None`` and should contain just the excluded genes data for each cell. Exclude all
cells whose sum of the excluded data divided by the total data is more than the specified
threshold.
"""
assert (max_excluded_genes_fraction is None) == (excluded_adata is None)
total_of_cells = ut.get_o_numpy(adata, what, sum=True)
cells_mask = np.full(adata.n_obs, True, dtype="bool")
if min_cell_total is not None:
cells_mask = cells_mask & (total_of_cells >= min_cell_total)
if max_cell_total is not None:
cells_mask = cells_mask & (total_of_cells <= max_cell_total)
if excluded_adata is not None:
assert max_excluded_genes_fraction is not None
excluded_data = ut.get_vo_proper(excluded_adata, layout="row_major")
excluded_of_cells = ut.sum_per(excluded_data, per="row")
if np.min(total_of_cells) == 0:
total_of_cells = np.copy(total_of_cells)
total_of_cells[total_of_cells == 0] = 1
excluded_fraction = excluded_of_cells / total_of_cells
cells_mask = cells_mask & (excluded_fraction <=
max_excluded_genes_fraction)
if inplace:
ut.set_o_data(adata, "properly_sampled_cell", cells_mask)
return None
ut.log_return("properly_sampled_cell", cells_mask)
return ut.to_pandas_series(cells_mask, index=adata.obs_names)
def renormalize_query_by_atlas( # pylint: disable=too-many-statements,too-many-branches
what: str = "__x__",
*,
adata: AnnData,
qdata: AnnData,
var_annotations: Dict[str, Any],
layers: Dict[str, Any],
varp_annotations: Dict[str, Any],
) -> Optional[AnnData]:
"""
Add an ``ATLASNORM`` pseudo-gene to query metacells data to compensate for the query having filtered out many genes.
This renormalizes the gene fractions in the query to fit the atlas in case the query has aggressive filtered a
significant amount of genes.
**Input**
Annotated query ``qdata`` and atlas ``adata``, where the observations are cells and the variables are genes, where
``X`` is a per-variable-per-observation matrix or the name of a per-variable-per-observation annotation containing
such a matrix.
**Returns**
None if no normalization is needed (or possible). Otherwise, a copy of the query metacells data, with an additional
variable (gene) called ``ATLASNORM`` to the query data, such that the total number of UMIs for each query metacells
is as expected given the total number of UMIs of the genes common to the query and the atlas. This is skipped if the
query and the atlas have exactly the same list of genes, or if if the query already contains a high number of genes
missing from the atlas so that the total number of UMIs for the query metacells is already at least the expected
based on the common genes.
**Computation Parameters**
1. Computes how many UMIs should be added to each query metacell so that its (total UMIs / total common gene UMIs)
would be the same as the (total atlas UMIs / total atlas common UMIs). If this is zero (or negative), stop.
2. Add an ``ATLASNORM`` pseudo-gene to the query with the above amount of UMIs. For each per-variable (gene)
observation, add the value specified in ``var_annotations``, whose list of keys must cover the set of
per-variable annotations in the query data. For each per-observation-per-variable layer, add the value specified
in ``layers``, whose list of keys must cover the existing layers. For each per-variable-per-variable annotation,
add the value specified in ``varp_annotations``.
"""
for name in qdata.var.keys():
if "|" not in name and name not in var_annotations.keys():
raise RuntimeError(f"missing default value for variable annotation {name}")
for name in qdata.layers.keys():
if name not in layers.keys():
raise RuntimeError(f"missing default value for layer {name}")
for name in qdata.varp.keys():
if name not in varp_annotations.keys():
raise RuntimeError(f"missing default value for variable-variable {name}")
if list(qdata.var_names) == list(adata.var_names):
return None
query_genes_list = list(qdata.var_names)
atlas_genes_list = list(adata.var_names)
common_genes_list = list(sorted(set(qdata.var_names) & set(adata.var_names)))
query_gene_indices = np.array([query_genes_list.index(gene) for gene in common_genes_list])
atlas_gene_indices = np.array([atlas_genes_list.index(gene) for gene in common_genes_list])
common_qdata = ut.slice(qdata, name=".common", vars=query_gene_indices, track_var="full_index")
common_adata = ut.slice(adata, name=".common", vars=atlas_gene_indices, track_var="full_index")
assert list(common_qdata.var_names) == list(common_adata.var_names)
atlas_total_umis_per_metacell = ut.get_o_numpy(adata, what, sum=True)
atlas_common_umis_per_metacell = ut.get_o_numpy(common_adata, what, sum=True)
atlas_total_umis = np.sum(atlas_total_umis_per_metacell)
atlas_common_umis = np.sum(atlas_common_umis_per_metacell)
atlas_disjoint_umis_fraction = atlas_total_umis / atlas_common_umis - 1.0
ut.log_calc("atlas_total_umis", atlas_total_umis)
ut.log_calc("atlas_common_umis", atlas_common_umis)
ut.log_calc("atlas_disjoint_umis_fraction", atlas_disjoint_umis_fraction)
query_total_umis_per_metacell = ut.get_o_numpy(qdata, what, sum=True)
query_common_umis_per_metacell = ut.get_o_numpy(common_qdata, what, sum=True)
query_total_umis = np.sum(query_total_umis_per_metacell)
query_common_umis = np.sum(query_common_umis_per_metacell)
query_disjoint_umis_fraction = query_total_umis / query_common_umis - 1.0
ut.log_calc("query_total_umis", query_total_umis)
ut.log_calc("query_common_umis", query_common_umis)
ut.log_calc("query_disjoint_umis_fraction", query_disjoint_umis_fraction)
if query_disjoint_umis_fraction >= atlas_disjoint_umis_fraction:
return None
query_normalization_umis_fraction = atlas_disjoint_umis_fraction - query_disjoint_umis_fraction
ut.log_calc("query_normalization_umis_fraction", query_normalization_umis_fraction)
query_normalization_umis_per_metacell = query_common_umis_per_metacell * query_normalization_umis_fraction
_proper, dense, compressed = ut.to_proper_matrices(qdata.X)
if dense is None:
assert compressed is not None
dense = ut.to_numpy_matrix(compressed)
added = np.concatenate([dense, query_normalization_umis_per_metacell[:, np.newaxis]], axis=1)
if compressed is not None:
added = sp.csr_matrix(added)
assert added.shape[0] == qdata.shape[0]
assert added.shape[1] == qdata.shape[1] + 1
ndata = AnnData(added)
ndata.obs_names = qdata.obs_names
var_names = list(qdata.var_names)
var_names.append("ATLASNORM")
ndata.var_names = var_names
for name, value in qdata.uns.items():
ut.set_m_data(ndata, name, value)
for name, value in qdata.obs.items():
ut.set_o_data(ndata, name, value)
for name, value in qdata.obsp.items():
ut.set_oo_data(ndata, name, value)
for name in qdata.var.keys():
if "|" in name:
continue
value = ut.get_v_numpy(qdata, name)
value = np.append(value, [var_annotations[name]])
ut.set_v_data(ndata, name, value)
for name in qdata.layers.keys():
data = ut.get_vo_proper(qdata, name)
_proper, dense, compressed = ut.to_proper_matrices(data)
if dense is None:
assert compressed is not None
dense = ut.to_numpy_matrix(compressed)
values = np.full(qdata.n_obs, layers[name], dtype=dense.dtype)
added = np.concatenate([dense, values[:, np.newaxis]], axis=1)
if compressed is not None:
added = sp.csr_matrix(added)
ut.set_vo_data(ndata, name, added)
for name in qdata.varp.keys():
data = ut.get_vv_proper(qdata, name)
_proper, dense, compressed = ut.to_proper_matrices(data)
if dense is None:
assert compressed is not None
dense = ut.to_numpy_matrix(compressed)
values = np.full(qdata.n_vars, varp_annotations[name], dtype=dense.dtype)
added = np.concatenate([dense, values[:, np.newaxis]], axis=1)
values = np.full(qdata.n_vars + 1, varp_annotations[name], dtype=dense.dtype)
added = np.concatenate([added, values[:, np.newaxis]], axis=0)
if compressed is not None:
added = sp.csr_matrix(added)
ut.set_vv_data(ndata, name, added)
return ndata
def compute_inner_normalized_variance(
what: Union[str, ut.Matrix] = "__x__",
*,
compatible_size: Optional[str] = None,
downsample_min_samples: int = pr.downsample_min_samples,
downsample_min_cell_quantile: float = pr.downsample_min_cell_quantile,
downsample_max_cell_quantile: float = pr.downsample_max_cell_quantile,
min_gene_total: int = pr.quality_min_gene_total,
adata: AnnData,
gdata: AnnData,
group: Union[str, ut.Vector] = "metacell",
random_seed: int = pr.random_seed,
) -> None:
"""
Compute the inner normalized variance (variance / mean) for each gene in each group.
This is also known as the "index of dispersion" and can serve as a quality measure for
the groups. An ideal group would contain only cells with "the same" biological state
and all remaining inner variance would be due to technical sampling noise.
**Input**
Annotated ``adata``, where the observations are cells and the variables are genes, where
``what`` is a per-variable-per-observation matrix or the name of a per-variable-per-observation
annotation containing such a matrix.
In addition, ``gdata`` is assumed to have one observation for each group, and use the same
genes as ``adata``.
**Returns**
Sets the following in ``gdata``:
Per-Variable Per-Observation (Gene-Cell) Annotations
``inner_variance``
For each gene and group, the variance of the gene in the group.
``inner_normalized_variance``
For each gene and group, the normalized variance (variance over mean) of the
gene in the group.
**Computation Parameters**
For each group (metacell):
1. If ``compatible_size`` (default: {compatible_size}) is specified, it should be an
integer per-observation annotation of the groups, whose value is at most the
number of grouped cells in the group. Pick a random subset of the cells of
this size. If ``compatible_size`` is ``None``, use all the cells of the group.
2. Invoke :py:func:`metacells.tools.downsample.downsample_cells` to downsample the surviving
cells to the same total number of UMIs, using the ``downsample_min_samples`` (default:
{downsample_min_samples}), ``downsample_min_cell_quantile`` (default:
{downsample_min_cell_quantile}), ``downsample_max_cell_quantile`` (default:
{downsample_max_cell_quantile}) and the ``random_seed`` (default: {random_seed}).
3. Compute the normalized variance of each gene based on the downsampled data. Set the
result to ``nan`` for genes with less than ``min_gene_total`` (default: {min_gene_total}).
"""
cells_data = ut.get_vo_proper(adata, what, layout="row_major")
if compatible_size is not None:
compatible_size_of_groups: Optional[ut.NumpyVector] = ut.get_o_numpy(
gdata, compatible_size, formatter=ut.sizes_description)
else:
compatible_size_of_groups = None
group_of_cells = ut.get_o_numpy(adata,
group,
formatter=ut.groups_description)
groups_count = np.max(group_of_cells) + 1
assert groups_count > 0
assert gdata.n_obs == groups_count
variance_per_gene_per_group = np.full(gdata.shape, None, dtype="float32")
normalized_variance_per_gene_per_group = np.full(gdata.shape,
None,
dtype="float32")
for group_index in range(groups_count):
with ut.log_step(
"- group",
group_index,
formatter=lambda group_index: ut.progress_description(
groups_count, group_index, "group"),
):
if compatible_size_of_groups is not None:
compatible_size_of_group = compatible_size_of_groups[
group_index]
else:
compatible_size_of_group = None
_collect_group_data(
group_index,
group_of_cells=group_of_cells,
cells_data=cells_data,
compatible_size=compatible_size_of_group,
downsample_min_samples=downsample_min_samples,
downsample_min_cell_quantile=downsample_min_cell_quantile,
downsample_max_cell_quantile=downsample_max_cell_quantile,
min_gene_total=min_gene_total,
random_seed=random_seed,
variance_per_gene_per_group=variance_per_gene_per_group,
normalized_variance_per_gene_per_group=
normalized_variance_per_gene_per_group,
)
ut.set_vo_data(gdata, "inner_variance", variance_per_gene_per_group)
ut.set_vo_data(gdata, "inner_normalized_variance",
normalized_variance_per_gene_per_group)
def compute_groups_self_consistency(
adata: AnnData,
what: Union[str, ut.Matrix] = "__x__",
*,
group: str = "metacell",
genes_mask: Optional[ut.NumpyVector] = None,
self_similarity_log_data: bool = pr.self_similarity_log_data,
self_similarity_value_normalization: float = pr.
self_similarity_value_normalization,
self_similarity_method: str = pr.self_similarity_method,
reproducible: bool = pr.reproducible,
logistics_location: float = pr.logistics_location,
logistics_slope: float = pr.logistics_slope,
) -> ut.NumpyVector:
"""
Compute the self consistency (similarity between two halves) of some groups.
**Input**
The input annotated ``adata`` is expected to contain a per-observation annotation named
``group`` (default: {group}) which identifies the group (metacells) each observation (cell)
belongs to, and ``half_<group>`` which identifies the half-group each observation belongs
to (e.g. as computed by :py:func:`split_groups`). Specifically, the indices of the halves
of group index ``i`` are ``i`` and ``i + groups_count``.
**Returns**
A Numpy vector holding, for each group, the similarity between its two halves.
**Computation Parameters**
1. For each group, compute the sum of values in each half and normalize it to fractions (sum of 1).
2. If ``genes_mask`` is specified, select only the genes specified in it. Note the sum of the
fractions of the genes of each group in the result will be less than or equal to 1.
3. If ``self_similarity_log_data`` (default: {self_similarity_log_data}), log2 the values using
``self_similarity_value_normalization`` (default: {self_similarity_value_normalization}).
4. Compute the ``self_similarity_method`` (default: {self_similarity_method}) between the two
halves. If this is the ``logistics`` similarity, then this will use ``logistics_location``
(default: {logistics_location}) and ``logistics_slope`` (default: {logistics_slope}). If this
is ``pearson``, and if ``reproducible`` (default: {reproducible}) is ``True``, a slower
(still parallel) but reproducible algorithm will be used to compute Pearson correlations.
"""
hdata = tl.group_obs_data(adata,
what,
groups=f"half_{group}",
name=".halves")
assert hdata is not None
sum_of_halves = ut.get_o_numpy(hdata, f"{what}|sum")
halves_values = ut.to_numpy_matrix(
ut.get_vo_proper(hdata, what, layout="row_major"))
halves_data = ut.mustbe_numpy_matrix(
ut.scale_by(halves_values, sum_of_halves, by="row"))
if self_similarity_value_normalization > 0:
halves_data += self_similarity_value_normalization
if self_similarity_log_data:
halves_data = ut.log_data(halves_data, base=2)
if genes_mask is not None:
halves_data = halves_data[:, genes_mask]
assert hdata.n_obs % 2 == 0
groups_count = hdata.n_obs // 2
low_half_indices = np.arange(groups_count)
high_half_indices = low_half_indices + groups_count
low_half_data = halves_data[low_half_indices, :]
high_half_data = halves_data[high_half_indices, :]
assert self_similarity_method in ("logistics", "pearson")
if self_similarity_method == "logistics":
similarity = ut.pairs_logistics_rows(low_half_data,
high_half_data,
location=logistics_location,
slope=logistics_slope)
similarity *= -1
similarity += 1
else:
similarity = ut.pairs_corrcoef_rows(low_half_data,
high_half_data,
reproducible=reproducible)
return similarity
def compute_significant_projected_fold_factors(
adata: AnnData,
what: Union[str, ut.Matrix] = "__x__",
*,
total_umis: Optional[ut.Vector],
projected: Union[str, ut.Matrix] = "projected",
fold_normalization: float = pr.project_fold_normalization,
min_significant_gene_value: float = pr.project_min_significant_gene_value,
min_gene_fold_factor: float = pr.project_max_projection_fold_factor,
min_entry_fold_factor: float = pr.min_entry_project_fold_factor,
abs_folds: bool = pr.project_abs_folds,
) -> None:
"""
Compute the significant projected fold factors of genes for each query metacell.
This computes, for each metacell of the query, the fold factors between the actual query UMIs and the UMIs of the
projection of the metacell onto the atlas (see :py:func:`metacells.tools.project.project_query_onto_atlas`). The
result per-metacell-per-gene matrix is then made sparse by discarding too-low values (setting them to zero).
Ideally, this matrix should be "very" sparse. If it contains "too many" non-zero values, more genes need to
be ignored by the projection, or somehow corrected for batch effects prior to computing the projection.
**Input**
Annotated ``adata``, where the observations are query metacells and the variables are genes, where ``what`` is a
per-variable-per-observation matrix or the name of a per-variable-per-observation annotation containing such a
matrix.
In addition, the ``projected`` UMIs of each query metacells onto the atlas.
**Returns**
Sets the following in ``gdata``:
Per-Variable Per-Observation (Gene-Cell) Annotations
``projected_fold``
For each gene and query metacell, the fold factor of this gene between the query and its projection (unless
the value is too low to be of interest, in which case it will be zero).
**Computation Parameters**
1. For each group (metacell), for each gene, compute the gene's fold factor
log2((actual UMIs + ``fold_normalization``) / (expected UMIs + ``fold_normalization``)), similarly to
:py:func:`metacells.tools.project.project_query_onto_atlas` (the default ``fold_normalization`` is
{fold_normalization}).
2. Set the fold factor to zero for every case where the total UMIs in the query metacell and the projected image is
not at least ``min_significant_gene_value`` (default: {min_significant_gene_value}).
3. If the maximal fold factor for a gene (across all metacells) is below ``min_gene_fold_factor`` (default:
{min_gene_fold_factor}), then set all the gene's fold factors to zero (too low to be of interest).
4. Otherwise, for any metacell whose fold factor for the gene is less than ``min_entry_fold_factor`` (default:
{min_entry_fold_factor}), set the fold factor to zero (too low to be of interest). If ``abs_folds`` (default:
{abs_folds}), consider the absolute fold factors.
"""
assert 0 <= min_entry_fold_factor <= min_gene_fold_factor
assert fold_normalization >= 0
metacells_data = ut.get_vo_proper(adata, what, layout="row_major")
projected_data = ut.get_vo_proper(adata, projected, layout="row_major")
metacells_fractions = ut.fraction_by(metacells_data,
by="row",
sums=total_umis)
projected_fractions = ut.fraction_by(projected_data,
by="row",
sums=total_umis)
metacells_fractions += fold_normalization # type: ignore
projected_fractions += fold_normalization # type: ignore
dense_folds = metacells_fractions / projected_fractions # type: ignore
dense_folds = np.log2(dense_folds, out=dense_folds)
total_umis = ut.to_numpy_matrix(metacells_data +
projected_data) # type: ignore
insignificant_folds_mask = total_umis < min_significant_gene_value
ut.log_calc("insignificant entries", insignificant_folds_mask)
dense_folds[insignificant_folds_mask] = 0.0
significant_folds = significant_folds_matrix(dense_folds,
min_gene_fold_factor,
min_entry_fold_factor,
abs_folds)
ut.set_vo_data(adata, "projected_fold", significant_folds)
def downsample_cells(
adata: AnnData,
what: Union[str, ut.Matrix] = "__x__",
*,
downsample_min_cell_quantile: float = pr.downsample_min_cell_quantile,
downsample_min_samples: float = pr.downsample_min_samples,
downsample_max_cell_quantile: float = pr.downsample_max_cell_quantile,
random_seed: int = pr.random_seed,
inplace: bool = True,
) -> Optional[ut.PandasFrame]:
"""
Downsample the values of ``what`` (default: {what}) data.
Downsampling is an effective way to get the same number of samples in multiple cells
(that is, the same number of total UMIs in multiple cells), and serves as an alternative to
normalization (e.g., working with UMI fractions instead of raw UMI counts).
Downsampling is especially important when computing correlations between cells. When there is
high variance between the total UMI count in different cells, then normalization will return
higher correlation values between cells with a higher total UMI count, which will result in an
inflated estimation of their similarity to other cells. Downsampling avoids this effect.
**Input**
Annotated ``adata``, where the observations are cells and the variables are genes, where
``what`` is a per-variable-per-observation matrix or the name of a per-variable-per-observation
annotation containing such a matrix.
**Returns**
Variable-Observation (Gene-Cell) Annotations
``downsampled``
The downsampled data where the total number of samples in each cell is at most
``samples``.
If ``inplace`` (default: {inplace}), this is written to the data, and the function returns
``None``. Otherwise this is returned as a pandas data frame (indexed by the cell and gene
names).
**Computation Parameters**
1. Compute the total samples in each cell.
2. Decide on the value to downsample to. We would like all cells to end up with at least some
reasonable number of samples (total UMIs) ``downsample_min_samples`` (default:
{downsample_min_samples}). We'd also like all (most) cells to end up with the highest
reasonable downsampled total number of samples, so if possible we increase the number of
samples, as long as at most ``downsample_min_cell_quantile`` (default:
{downsample_min_cell_quantile}) cells will have lower number of samples. We'd also like all
(most) cells to end up with the same downsampled total number of samples, so if we have to we
decrease the number of samples to ensure at most ``downsample_max_cell_quantile`` (default:
{downsample_max_cell_quantile}) cells will have a lower number of samples.
3. Downsample each cell so that it has at most the selected number of samples. Use the
``random_seed`` to allow making this replicable.
"""
total_per_cell = ut.get_o_numpy(adata, what, sum=True)
samples = int(
round(
min(
max(downsample_min_samples,
np.quantile(total_per_cell, downsample_min_cell_quantile)),
np.quantile(total_per_cell, downsample_max_cell_quantile),
)))
ut.log_calc("samples", samples)
data = ut.get_vo_proper(adata, what, layout="row_major")
assert ut.shaped_dtype(data) == "float32"
downsampled = ut.downsample_matrix(data,
per="row",
samples=samples,
random_seed=random_seed)
if inplace:
ut.set_vo_data(adata, "downsampled", downsampled)
return None
return ut.to_pandas_frame(downsampled,
index=adata.obs_names,
columns=adata.var_names)
def compute_direct_metacells( # pylint: disable=too-many-statements,too-many-branches
adata: AnnData,
what: Union[str, ut.Matrix] = "__x__",
*,
feature_downsample_min_samples: int = pr.feature_downsample_min_samples,
feature_downsample_min_cell_quantile: float = pr.feature_downsample_min_cell_quantile,
feature_downsample_max_cell_quantile: float = pr.feature_downsample_max_cell_quantile,
feature_min_gene_total: Optional[int] = pr.feature_min_gene_total,
feature_min_gene_top3: Optional[int] = pr.feature_min_gene_top3,
feature_min_gene_relative_variance: Optional[float] = pr.feature_min_gene_relative_variance,
feature_gene_names: Optional[Collection[str]] = None,
feature_gene_patterns: Optional[Collection[Union[str, Pattern]]] = None,
forbidden_gene_names: Optional[Collection[str]] = None,
forbidden_gene_patterns: Optional[Collection[Union[str, Pattern]]] = None,
cells_similarity_value_normalization: float = pr.cells_similarity_value_normalization,
cells_similarity_log_data: bool = pr.cells_similarity_log_data,
cells_similarity_method: str = pr.cells_similarity_method,
target_metacell_size: float = pr.target_metacell_size,
max_cell_size: Optional[float] = pr.max_cell_size,
max_cell_size_factor: Optional[float] = pr.max_cell_size_factor,
cell_sizes: Optional[Union[str, ut.Vector]] = pr.cell_sizes,
knn_k: Optional[int] = pr.knn_k,
min_knn_k: Optional[int] = pr.min_knn_k,
knn_balanced_ranks_factor: float = pr.knn_balanced_ranks_factor,
knn_incoming_degree_factor: float = pr.knn_incoming_degree_factor,
knn_outgoing_degree_factor: float = pr.knn_outgoing_degree_factor,
candidates_cell_seeds: Optional[Union[str, ut.Vector]] = None,
min_seed_size_quantile: float = pr.min_seed_size_quantile,
max_seed_size_quantile: float = pr.max_seed_size_quantile,
candidates_cooldown_pass: float = pr.cooldown_pass,
candidates_cooldown_node: float = pr.cooldown_node,
candidates_cooldown_phase: float = pr.cooldown_phase,
candidates_min_split_size_factor: Optional[float] = pr.candidates_min_split_size_factor,
candidates_max_merge_size_factor: Optional[float] = pr.candidates_max_merge_size_factor,
candidates_min_metacell_cells: Optional[int] = pr.min_metacell_cells,
candidates_max_split_min_cut_strength: Optional[float] = pr.max_split_min_cut_strength,
candidates_min_cut_seed_cells: Optional[int] = pr.min_cut_seed_cells,
must_complete_cover: bool = False,
deviants_min_gene_fold_factor: float = pr.deviants_min_gene_fold_factor,
deviants_abs_folds: bool = pr.deviants_abs_folds,
deviants_max_gene_fraction: Optional[float] = pr.deviants_max_gene_fraction,
deviants_max_cell_fraction: Optional[float] = pr.deviants_max_cell_fraction,
dissolve_min_robust_size_factor: Optional[float] = pr.dissolve_min_robust_size_factor,
dissolve_min_convincing_size_factor: Optional[float] = pr.dissolve_min_convincing_size_factor,
dissolve_min_convincing_gene_fold_factor: float = pr.dissolve_min_convincing_gene_fold_factor,
dissolve_min_metacell_cells: int = pr.dissolve_min_metacell_cells,
random_seed: int = pr.random_seed,
) -> AnnData:
"""
Directly compute metacells using ``what`` (default: {what}) data.
This directly computes the metacells on the whole data. Like any method that directly looks at
the whole data at once, the amount of CPU and memory needed becomes unreasonable when the data
size grows. Above O(10,000) you are much better off using the divide-and-conquer method.
.. note::
The current implementation is naive in that it computes the full dense N^2 correlation
matrix, and only then extracts the sparse graph out of it. We actually need two copies where
each requires 4 bytes per entry, so for O(100,000) cells, we have storage of
O(100,000,000,000). In addition, the implementation is serial for the graph clustering
phases.
It is possible to mitigate this by fusing the correlations phase and the graph generation
phase, parallelizing the result, and also (somehow) parallelizing the graph clustering
phase. This might increase the "reasonable" size for the direct approach to O(100,000).
We have decided not to invest in this direction since it won't allow us to push the size to
O(1,000,000) and above. Instead we provide the divide-and-conquer method, which easily
scales to O(1,000,000) on a single multi-core server, and to "unlimited" size if we further
enhance the implementation to use a distributed compute cluster of such servers.
.. todo::
Should :py:func:`compute_direct_metacells` avoid computing the graph and partition it for a
very small number of cells?
**Input**
The presumably "clean" annotated ``adata``, where the observations are cells and the variables
are genes, where ``what`` is a per-variable-per-observation matrix or the name of a
per-variable-per-observation annotation containing such a matrix.
**Returns**
Sets the following annotations in ``adata``:
Variable (Gene) Annotations
``high_total_gene``
A boolean mask of genes with "high" expression level.
``high_relative_variance_gene``
A boolean mask of genes with "high" normalized variance, relative to other genes with a
similar expression level.
``forbidden_gene``
A boolean mask of genes which are forbidden from being chosen as "feature" genes based
on their name.
``feature_gene``
A boolean mask of the "feature" genes.
``gene_deviant_votes``
The number of cells each gene marked as deviant (if zero, the gene did not mark any cell
as deviant). This will be zero for non-"feature" genes.
Observation (Cell) Annotations
``seed``
The index of the seed metacell each cell was assigned to to. This is ``-1`` for
non-"clean" cells.
``candidate``
The index of the candidate metacell each cell was assigned to to. This is ``-1`` for
non-"clean" cells.
``cell_deviant_votes``
The number of genes that were the reason the cell was marked as deviant (if zero, the
cell is not deviant).
``dissolved``
A boolean mask of the cells contained in a dissolved metacell.
``metacell``
The integer index of the metacell each cell belongs to. The metacells are in no
particular order. Cells with no metacell assignment ("outliers") are given a metacell
index of ``-1``.
``outlier``
A boolean mask of the cells contained in no metacell.
**Computation Parameters**
1. Invoke :py:func:`metacells.pipeline.feature.extract_feature_data` to extract "feature" data
from the clean data, using the
``feature_downsample_min_samples`` (default: {feature_downsample_min_samples}),
``feature_downsample_min_cell_quantile`` (default: {feature_downsample_min_cell_quantile}),
``feature_downsample_max_cell_quantile`` (default: {feature_downsample_max_cell_quantile}),
``feature_min_gene_total`` (default: {feature_min_gene_total}), ``feature_min_gene_top3``
(default: {feature_min_gene_top3}), ``feature_min_gene_relative_variance`` (default:
{feature_min_gene_relative_variance}), ``feature_gene_names`` (default:
{feature_gene_names}), ``feature_gene_patterns`` (default: {feature_gene_patterns}),
``forbidden_gene_names`` (default: {forbidden_gene_names}), ``forbidden_gene_patterns``
(default: {forbidden_gene_patterns}) and ``random_seed`` (default: {random_seed}) to make
this replicable.
2. Compute the fractions of each variable in each cell, and add the
``cells_similarity_value_normalization`` (default: {cells_similarity_value_normalization}) to
it.
3. If ``cells_similarity_log_data`` (default: {cells_similarity_log_data}), invoke the
:py:func:`metacells.utilities.computation.log_data` function to compute the log (base 2) of
the data.
4. Invoke :py:func:`metacells.tools.similarity.compute_obs_obs_similarity` to compute the
similarity between each pair of cells, using the
``cells_similarity_method`` (default: {cells_similarity_method}).
5. Invoke :py:func:`metacells.pipeline.collect.compute_effective_cell_sizes` using
``max_cell_size`` (default: {max_cell_size}), ``max_cell_size_factor`` (default:
{max_cell_size_factor}) and ``cell_sizes`` (default: {cell_sizes}) to get the effective cell
sizes to use.
5. Invoke :py:func:`metacells.tools.knn_graph.compute_obs_obs_knn_graph` to compute a
K-Nearest-Neighbors graph, using the
``knn_balanced_ranks_factor`` (default: {knn_balanced_ranks_factor}),
``knn_incoming_degree_factor`` (default: {knn_incoming_degree_factor})
and
``knn_outgoing_degree_factor`` (default: {knn_outgoing_degree_factor}).
If ``knn_k`` (default: {knn_k}) is not specified, then it is
chosen to be the median number of cells required to reach the target metacell size,
but at least ``min_knn_k`` (default: {min_knn_k}).
6. Invoke :py:func:`metacells.tools.candidates.compute_candidate_metacells` to compute
the candidate metacells, using the
``candidates_cell_seeds`` (default: {candidates_cell_seeds}),
``min_seed_size_quantile`` (default: {min_seed_size_quantile}),
``max_seed_size_quantile`` (default: {max_seed_size_quantile}),
``candidates_cooldown_pass`` (default: {candidates_cooldown_pass}),
``candidates_cooldown_node`` (default: {candidates_cooldown_node}),
``candidates_cooldown_phase`` (default: {candidates_cooldown_phase}),
``candidates_min_split_size_factor`` (default: {candidates_min_split_size_factor}),
``candidates_max_merge_size_factor`` (default: {candidates_max_merge_size_factor}),
``candidates_min_metacell_cells`` (default: {candidates_min_metacell_cells}),
and
``random_seed`` (default: {random_seed})
to make this replicable. This tries to build metacells of the
``target_metacell_size`` (default: {target_metacell_size})
using the effective cell sizes.
7. Unless ``must_complete_cover`` (default: {must_complete_cover}), invoke
:py:func:`metacells.tools.deviants.find_deviant_cells` to remove deviants from the candidate
metacells, using the
``deviants_min_gene_fold_factor`` (default: {deviants_min_gene_fold_factor}),
``deviants_abs_folds`` (default: {deviants_abs_folds}),
``deviants_max_gene_fraction`` (default: {deviants_max_gene_fraction})
and
``deviants_max_cell_fraction`` (default: {deviants_max_cell_fraction}).
8. Unless ``must_complete_cover`` (default: {must_complete_cover}), invoke
:py:func:`metacells.tools.dissolve.dissolve_metacells` to dissolve small unconvincing
metacells, using the same
``target_metacell_size`` (default: {target_metacell_size}),
and the effective cell sizes
and the
``dissolve_min_robust_size_factor`` (default: {dissolve_min_robust_size_factor}),
``dissolve_min_convincing_size_factor`` (default: {dissolve_min_convincing_size_factor}),
``dissolve_min_convincing_gene_fold_factor`` (default: {dissolve_min_convincing_size_factor})
and
``dissolve_min_metacell_cells`` (default: ``dissolve_min_metacell_cells``).
"""
fdata = extract_feature_data(
adata,
what,
top_level=False,
downsample_min_samples=feature_downsample_min_samples,
downsample_min_cell_quantile=feature_downsample_min_cell_quantile,
downsample_max_cell_quantile=feature_downsample_max_cell_quantile,
min_gene_relative_variance=feature_min_gene_relative_variance,
min_gene_total=feature_min_gene_total,
min_gene_top3=feature_min_gene_top3,
forced_gene_names=feature_gene_names,
forced_gene_patterns=feature_gene_patterns,
forbidden_gene_names=forbidden_gene_names,
forbidden_gene_patterns=forbidden_gene_patterns,
random_seed=random_seed,
)
if fdata is None:
raise ValueError("Empty feature data, giving up")
effective_cell_sizes, max_cell_size, _cell_scale_factors = compute_effective_cell_sizes(
adata, max_cell_size=max_cell_size, max_cell_size_factor=max_cell_size_factor, cell_sizes=cell_sizes
)
ut.log_calc("effective_cell_sizes", effective_cell_sizes, formatter=ut.sizes_description)
if max_cell_size is not None:
if candidates_min_metacell_cells is not None:
target_metacell_size = max(target_metacell_size, max_cell_size * candidates_min_metacell_cells)
if dissolve_min_metacell_cells is not None:
target_metacell_size = max(target_metacell_size, max_cell_size * dissolve_min_metacell_cells)
if candidates_min_metacell_cells is not None or dissolve_min_metacell_cells is not None:
ut.log_calc("target_metacell_size", target_metacell_size)
data = ut.get_vo_proper(fdata, "downsampled", layout="row_major")
data = ut.to_numpy_matrix(data, copy=True)
if cells_similarity_value_normalization > 0:
data += cells_similarity_value_normalization
if cells_similarity_log_data:
data = ut.log_data(data, base=2)
if knn_k is None:
if effective_cell_sizes is None:
median_cell_size = 1.0
else:
median_cell_size = float(np.median(effective_cell_sizes))
knn_k = int(round(target_metacell_size / median_cell_size))
if min_knn_k is not None:
knn_k = max(knn_k, min_knn_k)
if knn_k == 0:
ut.log_calc("knn_k: 0 (too small, try single metacell)")
ut.set_o_data(fdata, "candidate", np.full(fdata.n_obs, 0, dtype="int32"), formatter=lambda _: "* <- 0")
elif knn_k >= fdata.n_obs:
ut.log_calc(f"knn_k: {knn_k} (too large, try single metacell)")
ut.set_o_data(fdata, "candidate", np.full(fdata.n_obs, 0, dtype="int32"), formatter=lambda _: "* <- 0")
else:
ut.log_calc("knn_k", knn_k)
tl.compute_obs_obs_similarity(fdata, data, method=cells_similarity_method, reproducible=(random_seed != 0))
tl.compute_obs_obs_knn_graph(
fdata,
k=knn_k,
balanced_ranks_factor=knn_balanced_ranks_factor,
incoming_degree_factor=knn_incoming_degree_factor,
outgoing_degree_factor=knn_outgoing_degree_factor,
)
tl.compute_candidate_metacells(
fdata,
target_metacell_size=target_metacell_size,
cell_sizes=effective_cell_sizes,
cell_seeds=candidates_cell_seeds,
min_seed_size_quantile=min_seed_size_quantile,
max_seed_size_quantile=max_seed_size_quantile,
cooldown_pass=candidates_cooldown_pass,
cooldown_node=candidates_cooldown_node,
cooldown_phase=candidates_cooldown_phase,
min_split_size_factor=candidates_min_split_size_factor,
max_merge_size_factor=candidates_max_merge_size_factor,
min_metacell_cells=candidates_min_metacell_cells,
max_split_min_cut_strength=candidates_max_split_min_cut_strength,
min_cut_seed_cells=candidates_min_cut_seed_cells,
must_complete_cover=must_complete_cover,
random_seed=random_seed,
)
ut.set_oo_data(adata, "obs_similarity", ut.get_oo_proper(fdata, "obs_similarity"))
ut.set_oo_data(adata, "obs_outgoing_weights", ut.get_oo_proper(fdata, "obs_outgoing_weights"))
seed_of_cells = ut.get_o_numpy(fdata, "seed", formatter=ut.groups_description)
ut.set_o_data(adata, "seed", seed_of_cells, formatter=ut.groups_description)
candidate_of_cells = ut.get_o_numpy(fdata, "candidate", formatter=ut.groups_description)
ut.set_o_data(adata, "candidate", candidate_of_cells, formatter=ut.groups_description)
if must_complete_cover:
assert np.min(candidate_of_cells) == 0
deviant_votes_of_genes = np.zeros(adata.n_vars, dtype="float32")
deviant_votes_of_cells = np.zeros(adata.n_obs, dtype="float32")
dissolved_of_cells = np.zeros(adata.n_obs, dtype="bool")
ut.set_v_data(adata, "gene_deviant_votes", deviant_votes_of_genes, formatter=ut.mask_description)
ut.set_o_data(adata, "cell_deviant_votes", deviant_votes_of_cells, formatter=ut.mask_description)
ut.set_o_data(adata, "dissolved", dissolved_of_cells, formatter=ut.mask_description)
ut.set_o_data(adata, "metacell", candidate_of_cells, formatter=ut.groups_description)
else:
tl.find_deviant_cells(
adata,
candidates=candidate_of_cells,
min_gene_fold_factor=deviants_min_gene_fold_factor,
abs_folds=deviants_abs_folds,
max_gene_fraction=deviants_max_gene_fraction,
max_cell_fraction=deviants_max_cell_fraction,
)
tl.dissolve_metacells(
adata,
candidates=candidate_of_cells,
target_metacell_size=target_metacell_size,
cell_sizes=effective_cell_sizes,
min_robust_size_factor=dissolve_min_robust_size_factor,
min_convincing_size_factor=dissolve_min_convincing_size_factor,
min_convincing_gene_fold_factor=dissolve_min_convincing_gene_fold_factor,
min_metacell_cells=dissolve_min_metacell_cells,
)
metacell_of_cells = ut.get_o_numpy(adata, "metacell", formatter=ut.groups_description)
outlier_of_cells = metacell_of_cells < 0
ut.set_o_data(adata, "outlier", outlier_of_cells, formatter=ut.mask_description)
return fdata
