pyCistopic_kidney_raw.py

import pycisTopic
import os
import pybiomart as pbm
from pycisTopic.qc import *


projdir = "./cisTopic"
tmp_dir = "./cisTopic/tmp/"

if not os.path.exists(tmp_dir):
    os.makedirs(tmp_dir)

fragments_dict = {'kideny': '/home/annac/datasets/kidney_10x/fragments.tsv.gz'}

# For mouse (mm39)
#dataset = pbm.Dataset(name='mmusculus_gene_ensembl',  host='http://www.ensembl.org')
# For mouse (mm10)
#dataset = pbm.Dataset(name='mmusculus_gene_ensembl',  host='http://nov2020.archive.ensembl.org/')
# For human (hg38)
dataset = pbm.Dataset(name='hsapiens_gene_ensembl',  host='http://www.ensembl.org')
# For human (hg19)
#dataset = pbm.Dataset(name='hsapiens_gene_ensembl',  host='http://grch37.ensembl.org/')
# For fly (dm6)
# dataset = pbm.Dataset(name='dmelanogaster_gene_ensembl',  host='http://www.ensembl.org')

annot = dataset.query(attributes=['chromosome_name', 'transcription_start_site', 'strand', 'external_gene_name', 'transcript_biotype'])
filter = annot['Chromosome/scaffold name'].str.contains('CHR|GL|JH|MT')
annot = annot[~filter]
annot['Chromosome/scaffold name'] = annot['Chromosome/scaffold name'].str.replace(r'(\b\S)', r'chr\1')
annot.columns=['Chromosome', 'Start', 'Strand', 'Gene', 'Transcript_type']
annot = annot[annot.Transcript_type == 'protein_coding']

# path_to_regions = [path_to_regions[key] for key in fragments_dict.keys()]

path_to_regions = '/home/annac/datasets/kidney_10x/H_Kidney_Cancer_Chromium_Nuc_Iso_vs_SaltyEZ_vs_ComplexTissueDP_atac_peaks.bed'

metadata_bc, profile_data_dict = compute_qc_stats(fragments_dict = fragments_dict,
                tss_annotation = annot,
                stats=['barcode_rank_plot', 'duplicate_rate', 'insert_size_distribution', 'profile_tss', 'frip'],
                label_list = None,
                path_to_regions = path_to_regions,
                n_cpu = 1,
                valid_bc = None,
                n_frag = 2000,
                n_bc = None,
                tss_flank_window = 2000,
                tss_window = 50,
                tss_minimum_signal_window = 100,
                tss_rolling_window = 10,
                remove_duplicates = True,
                _temp_dir = tmp_dir + 'ray_spill',
                use_polars = True)