def minimal_media(
manifest, model_folder, summarize=True, min_growth=0.1, threads=1
):
"""Calculate the minimal medium for a set of community models."""
samples = manifest.sample_id.unique()
paths = [
(
s,
path.join(
model_folder, manifest[manifest.sample_id == s].file.iloc[0]
),
)
for s in samples
]
args = [[s, p, min_growth] for s, p in paths]
results = workflow(_medium, args, threads)
if any(r is None for r in results):
raise OptimizationError(
"Could not find a growth medium that allows the specified "
"growth rate for all taxa in all samples :("
)
results = pd.concat(results, axis=0)
if summarize:
medium = results.groupby("reaction").flux.max().reset_index()
medium["metabolite"] = medium.reaction.str.replace("EX_", "")
return medium
def tradeoff(
manifest,
model_folder,
medium,
tradeoffs=np.arange(0.1, 1.0 + 1e-6, 0.1),
threads=1,
):
"""Run growth rate predictions for varying tradeoff values.
Parameters
----------
manifest : pandas.DataFrame
The manifest as returned by the `build` workflow.
model_folder : str
The folder in which to find the files mentioned in the manifest.
medium : pandas.DataFrame
A growth medium. Must have columns "reaction" and "flux" denoting
exchnage reactions and their respective maximum flux.
tradeoffs : array of floats in (0.0, 1.0]
An array of tradeoff vaues to be tested. One simulation without
a tradeoff (no cooperative tradeoff) will always be run additionally
and will have a tradeoff of "NaN".
threads : int >=1
The number of parallel workers to use when building models. As a
rule of thumb you will need around 1GB of RAM for each thread.
Returns
-------
pandas.DataFrame
The predicted growth rates.
"""
samples = manifest.sample_id.unique()
paths = {
s: path.join(model_folder,
manifest[manifest.sample_id == s].file.iloc[0])
for s in samples
}
if any(t < 0.0 or t > 1.0 for t in tradeoffs):
raise ValueError("tradeoff values must between 0 and 1 :(")
medium = process_medium(medium, samples)
args = [[p, tradeoffs, medium.flux[medium.sample_id == s]]
for s, p in paths.items()]
results = workflow(_tradeoff, args, threads)
if all(r is None for r in results):
raise OptimizationError(
"All numerical optimizations failed. This indicates a problem "
"with the solver or numerical instabilities. Check that you have "
"CPLEX or Gurobi installed. You may also increase the abundance "
"cutoff in `qiime micom build` to create simpler models or choose "
"a more permissive solver tolerance.")
results = pd.concat(results)
return results
def check_db_medium(model_db, medium, threads=1):
"""Complete a growth medium for all models in a database.
Arguments
---------
model_db : str
A pre-built model database. If ending in `.qza` must be a Qiime 2
artifact of type `MetabolicModels[JSON]`. Can also be a folder,
zip (must end in `.zip`) file or None if the taxonomy contains a
column `file`.
medium : pd.DataFrame
A growth medium. Must have columns "reaction" and "flux" denoting
exchange reactions and their respective maximum flux. Can not be sample
specific.
threads : int >=1
The number of parallel workers to use when building models. As a
rule of thumb you will need around 1GB of RAM for each thread.
Returns
-------
pd.DataFrame
Returns an annotated manifest file with a column `can_grow` that tells you
whether the model can grow on the (fixed) medium, and a column `growth_rate`
that gives the growth rate.
"""
medium = process_medium(medium, ["dummy"])
medium.index = medium.global_id
compressed = model_db.endswith(".qza") or model_db.endswith(".zip")
if compressed:
tdir = TemporaryDirectory(prefix="micom_")
if model_db.endswith(".qza"):
manifest = load_qiime_model_db(model_db, tdir.name)
elif model_db.endswith(".zip"):
manifest = load_zip_model_db(model_db, tdir.name)
else:
manifest = load_manifest(model_db)
rank = manifest["summary_rank"][0]
logger.info("Checking %d %s-level models on a medium with %d components." %
(manifest.shape[0], rank, len(medium)))
args = [(f, medium.flux) for f in manifest.file]
results = workflow(_grow, args, threads)
manifest["growth_rate"] = results
manifest["can_grow"] = manifest.growth_rate.notna() & (manifest.growth_rate
> 1e-6)
if compressed:
tdir.cleanup()
return manifest
def db_annotations(
model_db,
threads=1,
):
"""Get metabolite annotations from a model DB.
Arguments
---------
model_db : str
A pre-built model database. If ending in `.qza` must be a Qiime 2
artifact of type `MetabolicModels[JSON]`. Can also be a folder,
zip (must end in `.zip`) file or None if the taxonomy contains a
column `file`.
threads : int >=1
The number of parallel workers to use when building models. As a
rule of thumb you will need around 1GB of RAM for each thread.
Returns
-------
pd.DataFrame
Annotations for all exchanged metabolites.
"""
compressed = model_db.endswith(".qza") or model_db.endswith(".zip")
if compressed:
tdir = TemporaryDirectory(prefix="micom_")
if model_db.endswith(".qza"):
manifest = load_qiime_model_db(model_db, tdir.name)
elif model_db.endswith(".zip"):
manifest = load_zip_model_db(model_db, tdir.name)
else:
manifest = load_manifest(model_db)
rank = manifest["summary_rank"][0]
logger.info("Getting annotations from %d %s-level models ." %
(manifest.shape[0], rank))
args = manifest.file.tolist()
results = workflow(_annotate, args, threads)
anns = pd.concat(results).drop_duplicates()
if compressed:
tdir.cleanup()
return anns
def grow(
manifest,
model_folder,
medium,
tradeoff,
threads=1,
weights=None,
atol=None,
rtol=None
):
"""Simulate growth for a set of community models.
Parameters
----------
manifest : pandas.DataFrame
The manifest as returned by the `build` workflow.
model_folder : str
The folder in which to find the files mentioned in the manifest.
medium : pandas.DataFrame
A growth medium. Must have columns "reaction" and "flux" denoting
exchange reactions and their respective maximum flux.
tradeoff : float in (0.0, 1.0]
A tradeoff value. Can be chosen by running the `tradeoff` workflow or
by experince. Tradeoff values of 0.5 for metagenomcis data and 0.3 for
16S data seem to work well.
threads : int >=1
The number of parallel workers to use when building models. As a
rule of thumb you will need around 1GB of RAM for each thread.
weights : str
Used during the calculaton of the minimal import rates.
Will scale the fluxes by a weight factor. Can either be "mass" which will
scale by molecular mass, a single element which will scale by
the elemental content (for instance "C" to scale by carbon content).
If None every metabolite will receive the same weight.
Will be ignored if `minimize_components` is True.
atol : float
Absolute tolerance for the growth rates. If None will use the solver tolerance.
rtol : float
Relative tolerqance for the growth rates. If None will use the solver tolerance.
Returns
-------
GrowthResults
A named tuple containing the growth rates and exchange fluxes for all
samples/models.
"""
samples = manifest.sample_id.unique()
paths = {
s: path.join(
model_folder, manifest[manifest.sample_id == s].file.iloc[0])
for s in samples
}
medium = process_medium(medium, samples)
args = [
[p, tradeoff, medium.flux[medium.sample_id == s], weights, atol, rtol]
for s, p in paths.items()
]
results = workflow(_growth, args, threads)
if all([r is None for r in results]):
raise OptimizationError(
"All numerical optimizations failed. This indicates a problem "
"with the solver or numerical instabilities. Check that you have "
"CPLEX or Gurobi installed. You may also increase the abundance "
"cutoff to create simpler models."
)
growth = pd.concat(r["growth"] for r in results if r is not None)
growth = growth[growth.taxon != "medium"]
exchanges = pd.concat(r["exchanges"] for r in results)
exchanges["taxon"] = exchanges.index
exchanges = exchanges.melt(
id_vars=["taxon", "sample_id", "tolerance"],
var_name="reaction", value_name="flux"
).dropna(subset=["flux"])
abundance = growth[["taxon", "sample_id", "abundance"]]
exchanges = pd.merge(exchanges, abundance,
on=["taxon", "sample_id"], how="outer")
anns = pd.concat(r["annotations"] for r in results).drop_duplicates()
anns.index = anns.reaction
exchanges["metabolite"] = anns.loc[exchanges.reaction, "metabolite"].values
exchanges["direction"] = DIRECTION[
(exchanges.flux > 0.0).astype(int)
].values
exchanges = exchanges[exchanges.flux.abs() > exchanges.tolerance]
return GrowthResults(growth, exchanges, anns)
def fix_medium(
manifest,
model_folder,
medium,
min_growth=0.1,
max_import=1,
minimize_components=False,
summarize=True,
weights=None,
threads=1,
):
"""Augment a growth medium so all community members can grow in it.
Arguments
---------
manifest : pandas.DataFrame
The manifest as returned by the `build` workflow.
model_folder : str
The folder in which to find the files mentioned in the manifest.
medium : pandas.Series or pandas.DataFrame
A growth medium with exchange reaction IDs as index and positive
import fluxes as values. If a DataFrame needs columns `flux` and
`reaction`.
min_growth : positive float
The minimum biomass production required for growth.
max_import : positive float
The maximum import rate for added imports.
minimize_components : boolean
Whether to minimize the number of media components rather than the
total flux.
summarize: boolean
Whether to summarize the medium across all samples. If False will
return a medium for each sample.
weights : str
Will scale the fluxes by a weight factor. Can either be "mass" which will
scale by molecular mass, a single element which will scale by
the elemental content (for instance "C" to scale by carbon content).
If None every metabolite will receive the same weight.
Will be ignored if `minimize_components` is True.
threads: int
The number of processes to use.
Returns
-------
pandas.DataFrame
A new growth medium with the smallest amount of augmentations such
that all members of the community can grow in it.
"""
if not isinstance(medium, pd.DataFrame):
raise ValueError("`medium` must be a DataFrame.")
samples = manifest.sample_id.unique()
paths = {
s: path.join(model_folder,
manifest[manifest.sample_id == s].file.iloc[0])
for s in samples
}
medium = process_medium(medium, samples)
if medium.flux[medium.flux < 1e-6].any():
medium.loc[medium < 1e-6, "flux"] = 1e-6
logger.info(
"Some import rates were to small and were adjusted to 1e-6.")
args = [[
s, p, min_growth, max_import, minimize_components,
medium.flux[medium.sample_id == s], weights
] for s, p in paths.items()]
res = workflow(_fix_medium, args, n_jobs=threads, unit="model(s)")
if all(r is None for r in res):
raise OptimizationError(
"All optimizations failed. You may need to increase `max_import` "
"or lower the target growth rate.")
final = pd.concat(res)
if summarize:
final = (final.groupby(["reaction", "metabolite",
"description"]).flux.max().reset_index())
return final
def complete_db_medium(
model_db,
medium,
growth=0.001,
max_added_import=1,
minimize_components=False,
weights=None,
threads=1,
):
"""Complete a growth medium for all models in a database.
Arguments
---------
model_db : str
A pre-built model database. If ending in `.qza` must be a Qiime 2
artifact of type `MetabolicModels[JSON]`. Can also be a folder,
zip (must end in `.zip`) file or None if the taxonomy contains a
column `file`.
medium : pd.DataFrame
A growth medium. Must have columns "reaction" and "flux" denoting
exchange reactions and their respective maximum flux. Can not be sample
specific.
growth : positive float or pandas.Series
The minimum growth rate the model has to achieve with the (fixed) medium. If
a Series will have a minimum growth rate for each id/taxon in the model db.
max_added_import : positive float
Maximum import flux for each added additional import not included in the growth
medium. If positive will expand the medium with additional imports in order to
fulfill the growth objective.
minimize_components : boolean
Whether to minimize the number of components instead of the total
import flux. Might be more intuitive if set to True but may also be
slow to calculate.
weights : str
Will scale the fluxes by a weight factor. Can either be "mass" which will
scale by molecular mass, a single element which will scale by
the elemental content (for instance "C" to scale by carbon content).
If None every metabolite will receive the same weight.
Will be ignored if `minimize_components` is True.
threads : int >=1
The number of parallel workers to use when building models. As a
rule of thumb you will need around 1GB of RAM for each thread.
Returns
-------
tuple of (manifest, import fluxes)
Returns an annotated manifest file with a column `can_grow` that tells you
whether the model can grow on the (fixed) medium, and a column `added` that
gives the number of added imports apart from the ones in the medium.
"""
medium = process_medium(medium, ["dummy"])
medium.index = medium.global_id
compressed = model_db.endswith(".qza") or model_db.endswith(".zip")
if compressed:
tdir = TemporaryDirectory(prefix="micom_")
if model_db.endswith(".qza"):
manifest = load_qiime_model_db(model_db, tdir.name)
elif model_db.endswith(".zip"):
manifest = load_zip_model_db(model_db, tdir.name)
else:
manifest = load_manifest(model_db)
rank = manifest["summary_rank"][0]
logger.info("Checking %d %s-level models on a medium with %d components." %
(manifest.shape[0], rank, len(medium)))
if not isinstance(growth, pd.Series):
growth = pd.Series(growth, index=manifest.id)
manifest.index = manifest.id
args = [(
manifest.loc[i, "file"],
medium.flux,
growth[i],
max_added_import,
minimize_components,
weights,
) for i in manifest.index]
results = workflow(_try_complete, args, threads)
manifest["can_grow"] = [r[0] for r in results]
manifest["added"] = [r[1] for r in results]
imports = pd.DataFrame.from_records([r[2] for r in results]).fillna(0.0)
imports.index = manifest.id
if compressed:
tdir.cleanup()
return (manifest, imports)
def build(
taxonomy, model_db, out_folder, cutoff=0.0001, threads=1, solver=None,
):
"""Builds a series of community models.
This is a best-practice implementation of building community models
for several samples in parallel.
Parameters
----------
taxonomy : pandas.DataFrame
The taxonomy used for building the model. Must have at least the
columns "id" and "sample_id". This must also
contain at least a column with the same name as the rank used in
the model database. Thus, for a genus-level database you will need
a column `genus`. Additional taxa ranks can also be specified and
will be used to be more stringent in taxa matching.
Finally, the taxonomy should contain a column `abundance`. It will
be used to quantify each individual in the community. If absent,
MICOM will assume all individuals are present in the same amount.
model_db : str
A pre-built model database. If ending in `.qza` must be a Qiime 2
artifact of type `MetabolicModels[JSON]`. Can also be a folder,
zip (must end in `.zip`) file or None if the taxonomy contains a
column `file`.
out_folder : str
The built models and a manifest file will be written to this
folder.
cutoff : float in [0.0, 1.0]
Abundance cutoff. Taxa with a relative abundance smaller than this
will not be included in the model.
threads : int >=1
The number of parallel workers to use when building models. As a
rule of thumb you will need around 1GB of RAM for each thread.
solver : str
Name of the solver used for the linear and quadratic problems.
Returns
-------
pandas.DataFrame
The manifest for the built models. Contains taxa abundances,
build metrics and file basenames.
"""
os.makedirs(out_folder, exist_ok=True)
samples = taxonomy.sample_id.unique()
out_path = pd.Series(
{s: os.path.join(out_folder, s + ".pickle") for s in samples}
)
args = [
[s, taxonomy[taxonomy.sample_id == s], model_db,
out_path[s], cutoff, solver]
for s in samples
]
res = workflow(build_and_save, args, threads)
metrics = pd.concat(res)
taxonomy = (
taxonomy.groupby("sample_id").apply(_reduce_group)
.dropna(axis=1).reset_index(drop=True)
)
taxonomy = taxonomy.loc[:, ~taxonomy.columns.isin(_ranks)]
taxonomy["file"] = taxonomy.sample_id + ".pickle"
taxonomy = pd.merge(taxonomy, metrics, on="sample_id")
taxonomy.to_csv(os.path.join(out_folder, "manifest.csv"), index=False)
return taxonomy
def build_database(
manifest, out_path, rank="genus", threads=1, compress=None, progress=True
):
"""Create a model database from a set of SBML files.
Note
----
A manifest for the joined models will also be written to the output folder
as "manifest.csv". This may contain NA entries for additional columns
that had different values within the summarized models.
Parameters
----------
manifest : pandas.DataFrame
A manifest of SBML files containing their filepath as well as taxonomy.
Must contain the columns "file", "kingdom", "phylum", "class",
"order", "family", "genus", and "species". May contain additional
columns.
out_path : str
The directory where the joined models will be written.
threads : int >=1
The number of parallel workers to use when building models. As a
rule of thumb you will need around 1GB of RAM for each thread.
compress : bool
Whether to compress the output. Default is True if out_path ends with
".zip" otherwise no.
progress : bool
Whether to show a progress bar.
Returns
-------
pd.DataFrame
The manifest of the joined models. Will still contain information
from the original metadata.
"""
meta = manifest.copy()
meta.columns = meta.columns.str.lower()
compress = out_path.endswith(".zip")
if not REQ_FIELDS.isin(meta.columns).all():
raise ValueError(
"Metadata File needs to have the following "
"columns %s." % ", ".join(REQ_FIELDS)
)
bad = meta.file.apply(lambda x: not os.path.exists(x))
if any(bad):
raise ValueError(
"The following models are in the manifest but do "
"not exist at the specified path: %s" % meta.file[bad]
)
meta = meta.groupby(rank).apply(_reduce_group).reset_index(drop=True)
logger.info("Building %d models on rank `%s`." % (meta.shape[0], rank))
meta.index = meta[rank].str.replace("[^\\w\\_]", "_")
meta["id"] = meta.index
meta["summary_rank"] = rank
if compress:
with TemporaryDirectory(prefix="micom_") as tdir:
args = [
(tid, row, os.path.join(tdir, "%s.json" % tid))
for tid, row in meta.iterrows()
]
workflow(_summarize_models, args, threads)
meta.file = meta.index + ".json"
meta.to_csv(os.path.join(tdir, "manifest.csv"), index=False)
with ZipFile(out_path, "w") as zf:
[zf.write(a[2], os.path.basename(a[2])) for a in args]
zf.write(os.path.join(tdir, "manifest.csv"), "manifest.csv")
else:
os.makedirs(out_path, exist_ok=True)
args = [
(tid, row, os.path.join(out_path, "%s.json" % tid))
for tid, row in meta.iterrows()
]
workflow(_summarize_models, args, threads)
meta.file = meta.index + ".json"
meta.to_csv(os.path.join(out_path, "manifest.csv"), index=False)
return meta
