"window": '--window '+str(args.window),
"duo_txt": str(duo_txt),
"thread_str": '--thread '+str(args.threads),
"seed_str": '--seed '+str(args.shape_seed),
"outmax": '--output-max '+str(outstem)+'.phased.haps '+str(outstem)+'.phased.sample',
"shapelog": str(outstem)+'.shape.resub_'+str(num_chr)+'.log',
"cbopen":'{{',
"cbclose":'}}',
}
shape_cmd = cmd_templ.format(**jobdict)
# submit
jobres = send_job(jobname='shape.'+str(outdot)+'.resub_'+str(num_chr),
cmd=shape_cmd,
logname='shape.'+str(outdot)+'.resub_'+str(num_chr)+'.sub.'+str(clust_conf['log_task_id'])+'.log',
mem=int(args.mem_req)*1000,
walltime=30,
njobs=int(num_chr),
threads=args.threads,
sleep=args.sleep)
print 'Pre-phasing jobs re-submitted for %d chromosomes.\n' % num_chr
# put this job back in the queue
print '\n...Replacing this imputation job in the queue...'
os.chdir(wd)
imp_log = 'imp_chunks.'+str(outdot)+'.sub.log'
send_job(jobname='imp.chunks.'+str(outdot),
]
print ' '.join(shape_call) + '\n'
# setup naming from task index
configs = read_conf(os.environ['HOME'] + '/picopili.conf')
clust_confdir = os.path.dirname(str(rp_bin)) + '/cluster_templates/'
clust_conf = read_conf(clust_confdir + str(configs['cluster'] + '.conf'))
task_id = str(clust_conf['log_task_id'])
# submit
jobres = send_job(jobname='shape.' + str(outdot),
cmd=' '.join(shape_call),
logname='shape.' + str(outdot) + '.chr' + task_id +
'.sub.log',
mem=int(args.mem_req) * 1000,
walltime=30,
njobs=22,
threads=int(args.threads),
sleep=str(args.sleep))
###
# submit next imputation task
###
if args.full_pipe:
######################
print '\n...Queuing best-guess script...'
######################
os.chdir(wd)
next_call = str(rp_bin) + '/imp2_rel.py ' + ' '.join(sys.argv[1:])
job_dict['cfile'] = tmp_chunk_file_name
# re-save new settings (primarily to track updating mem and walltime)
save_job(jfile=orig_job_conf,
cmd_templ=cmd_templ,
job_dict=job_dict,
sendjob_dict=sendjob_dict)
# submit
imp_cmd = cmd_templ.format(**job_dict)
jobres = send_job(jobname=sendjob_dict['jobname'],
cmd=imp_cmd,
logname=sendjob_dict['logname'],
mem=sendjob_dict['mem'],
walltime=sendjob_dict['walltime'],
njobs=sendjob_dict['njobs'],
sleep=sendjob_dict['sleep'],
forcearray=True)
print 'GWAS jobs resubmitted for %d chunks.\n' % nummiss
print '\n...Replacing this best-guess job in the queue...'
os.chdir(wd)
bg_log = 'bg.' + str(outdot) + '.resub_' + str(nummiss) + '.sub.log'
# TODO: consider queue/mem for agg
send_job(
jobname='bg.chunks.' + str(outdot),
cmd=' '.join(sys.argv[:]),
#############
# TODO: here
#############
print '\n...Submitting first task...'
#############
rp_bin = os.path.dirname(os.path.realpath(__file__))
next_call = str(rp_bin) + '/shape_rel.py ' + ' '.join(
sys.argv[1:]) + ' --full-pipe'
shape_log = 'shape.' + str(outdot) + '.sub.log'
# TODO: consider queue/mem
send_job(jobname='shape.' + str(outdot),
cmd=next_call,
logname=shape_log,
mem=int(args.mem_req * 1000),
walltime=30,
sleep=args.sleep)
# finish
print '\n############'
print '\n'
print 'SUCCESS!'
print 'All jobs submitted.\n'
exit(0)
# eof
job_dict['cfile'] = tmp_chunk_file_name
# re-save new settings (primarily to track updating mem and walltime)
save_job(jfile=orig_job_conf, cmd_templ=cmd_templ, job_dict=job_dict, sendjob_dict=sendjob_dict)
# submit
gwas_cmd = cmd_templ.format(**job_dict)
jobres = send_job(jobname=sendjob_dict['jobname'],
cmd=gwas_cmd,
logname=sendjob_dict['logname'],
mem=sendjob_dict['mem'],
walltime=sendjob_dict['walltime'],
njobs=sendjob_dict['njobs'],
maxpar=sendjob_dict['maxpar'],
sleep=sendjob_dict['sleep'],
forcearray=True)
print 'GWAS jobs resubmitted for %d chunks.\n' % nummiss
print '\n...Replacing this agg job in the queue...'
# TODO: adjust memory setting here
agg_log = 'agg.'+str(outdot)+'.resub_'+str(nummiss)+'.sub.log'
send_job(jobname='agg_'+str(outdot),
clust_dict['mem'] = max(4000,args.plink_mem)
clust_dict['walltime'] = 2
clust_dict['njobs'] = int(nchunk)
clust_dict['maxpar'] = 200
clust_dict['sleep'] = args.sleep
save_job(jfile=job_store_file, cmd_templ=gwas_templ, job_dict=jobdict, sendjob_dict=clust_dict)
# submit job
gwas_cmd = gwas_templ.format(**jobdict)
jobres = send_job(jobname='gwas.chunks.'+str(outdot),
cmd=gwas_cmd,
logname=str('gwas.chunks.'+str(outdot)+'.'+str(clust_conf['log_task_id'])+'.sub.log'),
mem=max(4000,args.plink_mem),
walltime=2,
njobs=int(nchunk),
maxpar=200,
sleep=args.sleep)
print 'GWAS jobs submitted for %d chunks.\n' % nchunk
######################
print '\n...Preparing meta-data for aggregation...'
# - create .frq file for aggregation script
# TODO: any prep for info score file?
######################
if args.model == 'linear':
'strict_qc.py', '--bfile', args.bfile, '--out', args.out, clean_txt,
'--mind-th',
str(args.mind_th), '--maf-th',
str(args.maf_th), '--hwe-th',
str(args.hwe_th), '--miss-th',
str(args.miss_th), '--ld-th',
str(args.ld_th), '--ld-wind',
str(args.ld_wind), '--plink-mem',
str(args.plink_mem), mhc_txt, chr8inv_txt, ldregions_txt, indel_txt,
strandambi_txt, allchr_txt
])
jobres = send_job(jobname=str('strictqc_' + args.out),
arrayfile=None,
cmd=str(strictqc_call),
logname=str('strictqc_' + args.out + '.sub.log'),
mem=int(args.plink_mem),
walltime=2,
sleep=0,
testonly=args.test_sub)
#####
# submit imus pca
print '\n...Submitting IMUS PCA job...'
#####
imuspca_call = ' '.join([
'imus_pca.py', '--bfile',
str(args.out + '.strictqc.pruned'), '--out', args.out, clean_txt,
'--rel-th',
str(args.rel_th), '--npcs',
str(args.npcs), plotall_txt, '--pcadir',
