def run(args):
if os.path.exists(args.model_output_path):
logging.info("DB already there, delete it if you want it done again")
return
folder = os.path.split(args.model_output_path)[0]
if not os.path.exists(folder):
os.makedirs(folder)
callback = callback_for_args(args)
PB8KFileInfo.parse_input_file(args.model_output_path, args.pb8k_input_path, args.gencode_input_path, callback)
def row_from_comps(comps, TF, var, N):
snp = comps[TF.SNPName]
if not snp in var:
return None
chr = "chr"+comps[TF.SNPChr]
pos = comps[TF.SNPChrPos]
reference_allele, effect_allele = PB8KFileInfo.alleles(TF, comps)
zscore = float(comps[TF.OverallZScore])
se = math.sqrt(float(var[snp])/N)
beta = zscore*se
OR = str(math.exp(beta))
#check ldpred coord_genotype parse_sum_stats_standard() for what is a1 and a2
p = str(stats.norm.sf(abs(zscore)) * 2)
row = (chr, snp, reference_allele, effect_allele, pos, OR, p,)
return row
def row_from_comps(gene, comps, TF):
snp = comps[TF.SNPName]
reference_allele, effect_allele = PB8KFileInfo.alleles(TF, comps)
zscore = comps[TF.OverallZScore]
row = (snp, gene, zscore, reference_allele, effect_allele,)
return row