def preprocess_dataframe(df: pd.DataFrame, params):
subframes = []
chain_dups_to_process = ("1", "2") if params.import_dual_chains is True else ("1")
for chain in params.receptor_chains.value:
for chain_dup in chain_dups_to_process:
subframe_dict = {"cell_ids": df["Clonotype ID"],
"sequence_aas": df[f"Chain: {chain} ({chain_dup})"],
"v_genes": df[f"{chain} - V gene ({chain_dup})"],
"j_genes": df[f"{chain} - J gene ({chain_dup})"],
"chains": Chain.get_chain(chain).value}
if params.extra_columns_to_load is not None:
for extra_col in params.extra_columns_to_load:
subframe_dict[extra_col] = df[extra_col]
subframes.append(pd.DataFrame(subframe_dict))
df = pd.concat(subframes, axis=0)
df.dropna(subset=["sequence_aas", "v_genes", "j_genes"], inplace=True)
df.reset_index(drop=True, inplace=True)
if params.import_all_gene_combinations:
df = IRISImport.import_all_gene_combinations(df)
else:
for gene_column in ("v_genes", "j_genes"):
processed_names = [IRISImport._load_gene(rn.choice(raw_v_string.split(" | "))) for raw_v_string in df[gene_column]]
df[gene_column] = processed_names
ImportHelper.drop_empty_sequences(df, params.import_empty_aa_sequences, params.import_empty_nt_sequences)
return df
def preprocess_dataframe(df: pd.DataFrame, params: DatasetImportParams):
df["frame_types"] = SequenceFrameType.IN.name
ImportHelper.junction_to_cdr3(df, params.region_type)
if not params.is_repertoire and params.paired:
n_single_chains = sum(df["sequence_identifiers"] == "0")
if n_single_chains > 0:
df.drop(df.loc[df["sequence_identifiers"] == "0"].index, inplace=True)
warnings.warn(f"VDJdbImport: {n_single_chains} single chains were removed when trying to create a ReceptorDataset.\n"
f"To import all chains as a SequenceDataset, use paired = False")
else:
df.loc[df["sequence_identifiers"] == "0", "sequence_identifiers"] = None
if "chains" not in df.columns:
df.loc[:, "chains"] = ImportHelper.load_chains_from_genes(df)
df["receptor_identifiers"] = df["sequence_identifiers"]
df["sequence_identifiers"] = VDJdbImport.get_sequence_identifiers(df["sequence_identifiers"], df["chains"])
ImportHelper.update_gene_info(df)
ImportHelper.drop_empty_sequences(df, params.import_empty_aa_sequences, params.import_empty_nt_sequences)
ImportHelper.drop_illegal_character_sequences(df, params.import_illegal_characters)
df = VDJdbImport.extract_meta_columns(df, params)
return df
def preprocess_dataframe(dataframe: pd.DataFrame, params: DatasetImportParams):
dataframe.loc[:, "frame_types"] = dataframe.frame_types.str.upper()
frame_type_list = ImportHelper.prepare_frame_type_list(params)
dataframe = dataframe[dataframe["frame_types"].isin(frame_type_list)]
dataframe.loc[:, "region_types"] = params.region_type.name
if params.region_type == RegionType.IMGT_CDR3:
if "sequences" in dataframe.columns:
dataframe.loc[:, 'sequences'] = [y[(84 - 3 * len(x)): 78] if x is not None else None for x, y in zip(dataframe['sequence_aas'], dataframe['sequences'])]
dataframe.loc[:, 'sequence_aas'] = dataframe["sequence_aas"].str[1:-1]
elif "sequences" in dataframe.columns:
dataframe.loc[:, 'sequences'] = [y[(81 - 3 * len(x)): 81] if x is not None else None for x, y in zip(dataframe['sequence_aas'], dataframe['sequences'])]
dataframe = AdaptiveImportHelper.parse_adaptive_germline_to_imgt(dataframe, params.organism)
dataframe = ImportHelper.standardize_none_values(dataframe)
ImportHelper.drop_empty_sequences(dataframe, params.import_empty_aa_sequences, params.import_empty_nt_sequences)
ImportHelper.drop_illegal_character_sequences(dataframe, params.import_illegal_characters)
if "chains" in dataframe.columns:
dataframe.loc[:, "chains"] = ImportHelper.load_chains(dataframe)
else:
# loading from v_subgroups is preferred as sometimes v_genes is None when v_subgroups is defined
if "v_subgroups" in dataframe.columns:
dataframe.loc[:, "chains"] = ImportHelper.load_chains_from_column(dataframe, "v_subgroups")
else:
dataframe.loc[:, "chains"] = ImportHelper.load_chains_from_genes(dataframe)
return dataframe
def import_dataset(params: dict, dataset_name: str) -> Dataset:
iris_params = IRISImportParams.build_object(**params)
dataset = ImportHelper.load_dataset_if_exists(params, iris_params, dataset_name)
if dataset is None:
if iris_params.is_repertoire:
dataset = ImportHelper.import_repertoire_dataset(IRISImport, iris_params, dataset_name)
else:
dataset = IRISImport.load_sequence_dataset(params, dataset_name)
return dataset
def preprocess_dataframe(df: pd.DataFrame, params: DatasetImportParams):
ImportHelper.junction_to_cdr3(df, params.region_type)
ImportHelper.drop_empty_sequences(df, params.import_empty_aa_sequences,
params.import_empty_nt_sequences)
ImportHelper.drop_illegal_character_sequences(
df, params.import_illegal_characters)
ImportHelper.update_gene_info(df)
return df
def import_dataset(params, dataset_name: str) -> ReceptorDataset:
generic_params = DatasetImportParams.build_object(**params)
filenames = ImportHelper.get_sequence_filenames(
generic_params.path, dataset_name)
PathBuilder.build(generic_params.result_path, warn_if_exists=True)
dataset = SingleLineReceptorImport._import_from_files(
filenames, generic_params)
dataset.name = dataset_name
dataset.params = ImportHelper.extract_sequence_dataset_params(
params=generic_params)
PickleExporter.export(dataset, generic_params.result_path)
return dataset
def preprocess_dataframe(df: pd.DataFrame, params: DatasetImportParams):
"""
Function for preprocessing data from a dataframe containing AIRR data, such that:
- productive sequences, sequences with stop codons or out of frame sequences are filtered according to specification
- if RegionType is CDR3, the leading C and trailing W are removed from the sequence to match the CDR3 definition
- if no chain column was specified, the chain is extracted from the v gene name
- the allele information is removed from the V and J genes
"""
if "productive" in df.columns:
df["frame_types"] = SequenceFrameType.OUT.name
df.loc[df["productive"] == True,
"frame_types"] = SequenceFrameType.IN.name
else:
df["frame_types"] = None
if "vj_in_frame" in df.columns:
df.loc[df["vj_in_frame"] == True,
"frame_types"] = SequenceFrameType.IN.name
if "stop_codon" in df.columns:
df.loc[df["stop_codon"] == True,
"frame_types"] = SequenceFrameType.STOP.name
if "productive" in df.columns:
frame_type_list = ImportHelper.prepare_frame_type_list(params)
df = df[df["frame_types"].isin(frame_type_list)]
if params.region_type == RegionType.IMGT_CDR3:
if "sequence_aas" not in df.columns and "sequences" not in df.columns:
if "cdr3" in df.columns:
df.rename(columns={"cdr3": "sequences"}, inplace=True)
if "cdr3_aa" in df.columns:
df.rename(columns={"cdr3_aa": "sequence_aas"},
inplace=True)
df.loc[:, "region_types"] = params.region_type.name
elif "junction" in params.column_mapping or "junction_aa" in params.column_mapping:
ImportHelper.junction_to_cdr3(df, params.region_type)
# todo else: support "full_sequence" import through regiontype?
if "chains" not in df.columns:
df.loc[:, "chains"] = ImportHelper.load_chains_from_genes(df)
df = ImportHelper.update_gene_info(df)
ImportHelper.drop_empty_sequences(df, params.import_empty_aa_sequences,
params.import_empty_nt_sequences)
ImportHelper.drop_illegal_character_sequences(
df, params.import_illegal_characters)
return df
def load_sequence_dataset(params: dict, dataset_name: str) -> Dataset:
iris_params = IRISImportParams.build_object(**params)
filenames = ImportHelper.get_sequence_filenames(iris_params.path, dataset_name)
file_index = 0
dataset_filenames = []
for index, filename in enumerate(filenames):
items = IRISSequenceImport.import_items(filename, paired=iris_params.paired,
all_dual_chains=iris_params.import_dual_chains,
all_genes=iris_params.import_all_gene_combinations)
while len(items) > iris_params.sequence_file_size or (index == len(filenames) - 1 and len(items) > 0):
dataset_filenames.append(iris_params.result_path + "batch_{}.pickle".format(file_index))
ImportHelper.store_sequence_items(dataset_filenames, items, iris_params.sequence_file_size)
items = items[iris_params.sequence_file_size:]
file_index += 1
return ReceptorDataset(filenames=dataset_filenames, file_size=iris_params.sequence_file_size, name=dataset_name) if iris_params.paired \
else SequenceDataset(filenames=dataset_filenames, file_size=iris_params.sequence_file_size, name=dataset_name)
def import_sequence_dataset(params: dict, dataset_name: str) -> RepertoireDataset:
base_result_path = params['result_path'] + "tmp_airr/"
unzipped_path = base_result_path + "tmp_unzipped/"
IReceptorImport._unzip_files(params['path'], unzipped_path, unzip_metadata=False)
airr_params = copy.deepcopy(params)
airr_params["path"] = unzipped_path
dataset = ImportHelper.import_dataset(AIRRImport, airr_params, dataset_name)
shutil.rmtree(unzipped_path)
return dataset
def import_repertoire_dataset(params: dict, dataset_name: str) -> RepertoireDataset:
base_result_path = params['result_path'] + "tmp_airr/"
metadata_file_path = base_result_path + "metadata.csv"
IReceptorImport._create_airr_repertoiredataset(params['path'], base_result_path, metadata_file_path)
airr_params = copy.deepcopy(params)
airr_params["path"] = base_result_path
airr_params["metadata_file"] = metadata_file_path
dataset = ImportHelper.import_dataset(AIRRImport, airr_params, dataset_name)
shutil.rmtree(base_result_path)
return dataset
def preprocess_dataframe(df: pd.DataFrame, params: DatasetImportParams):
df["frame_types"] = None
df.loc[df["productive"].eq("True"),
"frame_types"] = SequenceFrameType.IN.name
allowed_productive_values = []
if params.import_productive:
allowed_productive_values.append("True")
if params.import_unproductive:
allowed_productive_values.append("False")
df = df[df.productive.isin(allowed_productive_values)]
ImportHelper.junction_to_cdr3(df, params.region_type)
ImportHelper.drop_empty_sequences(df, params.import_empty_aa_sequences,
params.import_empty_nt_sequences)
ImportHelper.drop_illegal_character_sequences(
df, params.import_illegal_characters)
ImportHelper.update_gene_info(df)
if "chains" not in df.columns:
df.loc[:, "chains"] = ImportHelper.load_chains_from_genes(df)
return df
def preprocess_dataframe(df: pd.DataFrame, params: DatasetImportParams):
"""
Function for loading the data from one MiXCR file, such that:
- for the given region (CDR3/full sequence), both nucleotide and amino acid sequence are loaded
- if the region is CDR3, it adapts the sequence to the definition of the CDR3 (IMGT junction vs IMGT CDR3)
- the chain for each sequence is extracted from the v gene name
- the genes are loaded from the top score for gene without allele info
Arguments:
metadata: the corresponding row from the metadata file with metadata such as subject_id, age, HLA or other info given there
params: DatasetImportParams object defining what to import and how to do it
Returns:
data frame corresponding to Repertoire.FIELDS and custom lists which can be used to create a Repertoire object
"""
df["sequence_aas"] = df[MiXCRImport.SEQUENCE_NAME_MAP[
params.region_type]["AA"]]
df["sequences"] = df[MiXCRImport.SEQUENCE_NAME_MAP[params.region_type]
["NT"]]
ImportHelper.junction_to_cdr3(df, params.region_type)
df["counts"] = df["counts"].astype(float).astype(int)
df["v_genes"] = MiXCRImport._load_genes(df, "v_genes")
df["j_genes"] = MiXCRImport._load_genes(df, "j_genes")
df["chains"] = ImportHelper.load_chains_from_genes(df)
ImportHelper.update_gene_info(df)
ImportHelper.drop_empty_sequences(df, params.import_empty_aa_sequences,
params.import_empty_nt_sequences)
ImportHelper.drop_illegal_character_sequences(
df, params.import_illegal_characters)
return df
def preprocess_dataframe(df: pd.DataFrame, params: DatasetImportParams):
if "sequences" not in df.columns and "sequence_aas" not in df.columns:
raise IOError(
"OLGAImport: Columns should contain at least 'sequences' or 'sequence_aas'."
)
if "counts" not in df.columns:
df["counts"] = 1
df["sequence_identifiers"] = None
ImportHelper.junction_to_cdr3(df, params.region_type)
ImportHelper.drop_empty_sequences(df, params.import_empty_aa_sequences,
params.import_empty_nt_sequences)
ImportHelper.drop_illegal_character_sequences(
df, params.import_illegal_characters)
if "chains" not in df.columns:
df.loc[:, "chains"] = ImportHelper.load_chains_from_genes(df)
return df
def prepare_reference(reference_params: dict, location: str, paired: bool):
ParameterValidator.assert_keys(list(reference_params.keys()),
["format", "params"], location,
"reference")
seq_import_params = reference_params[
"params"] if "params" in reference_params else {}
assert os.path.isfile(seq_import_params["path"]), f"{location}: the file {seq_import_params['path']} does not exist. " \
f"Specify the correct path under reference."
if "paired" in seq_import_params:
assert seq_import_params[
"paired"] == paired, f"{location}: paired must be {paired} for SequenceImport"
else:
seq_import_params["paired"] = paired
format_str = reference_params["format"]
if format_str == "IRIS": # todo refactor this when refactoring IRISSequenceImport
receptors = IRISSequenceImport.import_items(**seq_import_params)
else:
import_class = ReflectionHandler.get_class_by_name(
"{}Import".format(format_str))
params = DefaultParamsLoader.load(
EnvironmentSettings.default_params_path + "datasets/",
DefaultParamsLoader.convert_to_snake_case(format_str))
for key, value in seq_import_params.items():
params[key] = value
params["paired"] = paired
processed_params = DatasetImportParams.build_object(**params)
receptors = ImportHelper.import_items(
import_class, reference_params["params"]["path"],
processed_params)
return receptors
def preprocess_dataframe(df: pd.DataFrame, params: DatasetImportParams):
if "counts" not in df.columns:
df["counts"] = 1
df = df[df.anchors_found == "1"]
if not params.import_out_of_frame:
df = df[df.is_inframe == "1"]
df["sequence_aas"] = df["sequences"].apply(IGoRImport.translate_sequence)
if not params.import_with_stop_codon:
no_stop_codon = ["*" not in seq for seq in df.sequence_aas]
df = df[no_stop_codon]
ImportHelper.junction_to_cdr3(df, params.region_type)
# note: import_empty_aa_sequences is set to true here; since IGoR doesnt output aa, this parameter is insensible
ImportHelper.drop_empty_sequences(df, True, params.import_empty_nt_sequences)
ImportHelper.drop_illegal_character_sequences(df, params.import_illegal_characters)
# chain or at least receptorsequence?
return df
def import_receptors(df, params):
df["receptor_identifiers"] = df["cell_id"]
return ImportHelper.import_receptors(df, params)
def alternative_load_func(filename, params):
df = airr.load_rearrangement(filename)
df = ImportHelper.standardize_none_values(df)
df.dropna(axis="columns", how="all", inplace=True)
return df
def import_dataset(params: dict, dataset_name: str) -> Dataset:
return ImportHelper.import_dataset(AIRRImport, params, dataset_name)
def import_dataset(params: dict, dataset_name: str) -> Dataset:
return ImportHelper.import_dataset(ImmunoSEQSampleImport, params,
dataset_name)
def import_receptors(df, params):
return ImportHelper.import_receptors(df, params)
def import_receptors(df, params):
df["receptor_identifiers"] = df["sequence_identifiers"]
return ImportHelper.import_receptors(df, params)
def import_dataset(params: dict, dataset_name: str) -> Dataset:
return ImportHelper.import_dataset(ImmunoSEQRearrangementImport, params, dataset_name)
def import_dataset(params: dict, dataset_name: str) -> Dataset:
return ImportHelper.import_dataset(TenxGenomicsImport, params,
dataset_name)
