def filter_values(table: biom.Table, th: float) -> biom.Table: """Filter out low-abundance features within each sample in a table. """ def filter_otus(data, id_, md): bound = th if th > 1 else data.sum() * th data[data < bound] = 0 return data table.transform(filter_otus, axis='sample') table.remove_empty(axis='observation') return table
def scale_biom(table: biom.Table, scale: float): """Multiply table values by a constant factor in place. Parameters ---------- table : biom.Table BIOM table to multiply. scale : float Scale factor. """ table.transform(lambda data, id_, md: data * scale, inplace=True)
def divide_biom(table: biom.Table, sizes: dict): """Divide table values by feature sizes in place. Parameters ---------- table : biom.Table BIOM table to divide. sizes : dict Feature size dictionary. """ table.transform(lambda data, id_, md: data / sizes[id_], axis='observation', inplace=True)
def round_biom(table: biom.Table): """Round a BIOM table's data to integers and drop empty observations in place. Parameters ---------- table : biom.Table BIOM table to round. Notes ----- This function will not drop empty samples. """ f = np.vectorize(intize) table.transform(lambda data, id_, md: f(data), axis='observation') table.remove_empty(axis='observation')
def round_biom(table: biom.Table, digits=None): """Round a BIOM table's data and drop empty observations in place. Parameters ---------- table : biom.Table BIOM table to round. digits : int, optional Digits after the decimal point. Notes ----- This function will not drop empty samples. """ f = np.vectorize(partial(rounder, digits=digits)) table.transform(lambda data, id_, md: f(data), axis='observation') table.remove_empty(axis='observation')
def main(): option_parser, opts, args =\ parse_command_line_parameters(**script_info) otu_table = load_table(opts.input_otu_fp) ids_to_load = otu_table.ids(axis='observation') if(opts.input_count_fp is None): #precalc file has specific name (e.g. 16S_13_5_precalculated.tab.gz) precalc_file_name='_'.join(['16S',opts.gg_version,'precalculated.tab.gz']) input_count_table=join(get_picrust_project_dir(),'picrust','data',precalc_file_name) else: input_count_table=opts.input_count_fp if opts.verbose: print "Loading trait table: ", input_count_table ext=path.splitext(input_count_table)[1] if (ext == '.gz'): count_table_fh = gzip.open(input_count_table,'rb') else: count_table_fh = open(input_count_table,'U') if opts.load_precalc_file_in_biom: count_table = load_table(count_table_fh) else: count_table = convert_precalc_to_biom(count_table_fh, ids_to_load) #Need to only keep data relevant to our otu list ids=[] for x in otu_table.iter(axis='observation'): ids.append(str(x[1])) ob_id=count_table.ids(axis='observation')[0] filtered_otus=[] filtered_values=[] for x in ids: if count_table.exists(x, axis='sample'): filtered_otus.append(x) filtered_values.append(otu_table.data(x, axis='observation')) filtered_otu_table = Table(filtered_values, filtered_otus, otu_table.ids()) copy_numbers_filtered={} for x in filtered_otus: value = count_table.get_value_by_ids(ob_id,x) try: #data can be floats so round them and make them integers value = int(round(float(value))) except ValueError: raise ValueError,\ "Invalid type passed as copy number for OTU ID %s. Must be int-able." % (value) if value < 1: raise ValueError, "Copy numbers must be greater than or equal to 1." copy_numbers_filtered[x]={opts.metadata_identifer:value} filtered_otu_table.add_metadata(copy_numbers_filtered, axis='observation') def metadata_norm(v, i, md): return v / float(md[opts.metadata_identifer]) normalized_table = filtered_otu_table.transform(metadata_norm, axis='observation') #move Observation Metadata from original to filtered OTU table normalized_table = transfer_observation_metadata(otu_table, normalized_table, 'observation') make_output_dir_for_file(opts.output_otu_fp) write_biom_table(normalized_table, opts.output_otu_fp)
def main(): option_parser, opts, args = parse_command_line_parameters(**script_info) otu_table = load_table(opts.input_otu_fp) ids_to_load = otu_table.ids(axis="observation") if opts.input_count_fp is None: # precalc file has specific name (e.g. 16S_13_5_precalculated.tab.gz) precalc_file_name = "_".join(["16S", opts.gg_version, "precalculated.tab.gz"]) input_count_table = join(get_picrust_project_dir(), "picrust", "data", precalc_file_name) else: input_count_table = opts.input_count_fp if opts.verbose: print "Loading trait table: ", input_count_table ext = path.splitext(input_count_table)[1] if ext == ".gz": count_table_fh = gzip.open(input_count_table, "rb") else: count_table_fh = open(input_count_table, "U") if opts.load_precalc_file_in_biom: count_table = load_table(count_table_fh) else: count_table = convert_precalc_to_biom(count_table_fh, ids_to_load) # Need to only keep data relevant to our otu list ids = [] for x in otu_table.iter(axis="observation"): ids.append(str(x[1])) ob_id = count_table.ids(axis="observation")[0] filtered_otus = [] filtered_values = [] for x in ids: if count_table.exists(x, axis="sample"): filtered_otus.append(x) filtered_values.append(otu_table.data(x, axis="observation")) filtered_otu_table = Table(filtered_values, filtered_otus, otu_table.ids()) copy_numbers_filtered = {} for x in filtered_otus: value = count_table.get_value_by_ids(ob_id, x) try: # data can be floats so round them and make them integers value = int(round(float(value))) except ValueError: raise ValueError, "Invalid type passed as copy number for OTU ID %s. Must be int-able." % (value) if value < 1: raise ValueError, "Copy numbers must be greater than or equal to 1." copy_numbers_filtered[x] = {opts.metadata_identifer: value} filtered_otu_table.add_metadata(copy_numbers_filtered, axis="observation") def metadata_norm(v, i, md): return v / float(md[opts.metadata_identifer]) normalized_table = filtered_otu_table.transform(metadata_norm, axis="observation") # move Observation Metadata from original to filtered OTU table normalized_table = transfer_observation_metadata(otu_table, normalized_table, "observation") make_output_dir_for_file(opts.output_otu_fp) write_biom_table(normalized_table, opts.output_otu_fp)