##======
# read in simulated sequences
##======
# trees, recomb_sites, seg_sites, positions, seqs
sim = sim_process.read_one_sim(ms_f, sim_args['num_sites'], sim_args['num_samples'])
##======
# predict introgressed/non-introgressed tracts
##======
state_seq, probs, hmm, hmm_init, ps = \
sim_predict.predict_introgressed(sim, sim_args, predict_args, \
train=True, method='posterior', \
only_poly=only_poly)
state_seq_blocks = sim_process.convert_to_blocks(state_seq, \
predict_args['states'])
##======
# output
##======
if write_ps:
predict.write_positions(ps, ps_f, str(i), 'I')
# summary info about HMM (before training)
sim_predict.write_hmm_line(hmm_init, out_init_f, i==0)
# summary info about HMM (after training)
sim_predict.write_hmm_line(hmm, out_f, i==0)
# locations of introgression
#x = 20
#thresholds = [float(i) / x for i in range(0, x + 2)]
thresholds = [0, .00001, .00005, .0001, .0005, .001, .005, .01, .05, .1, .5, .6, .7, .8, .9, .99, 1, 1.1]
header = True
for threshold in thresholds:
print 'threshold:', threshold
all_stats = []
for rep in range(sim_args['num_reps']):
paths_t = {}
for ind in probs[rep]:
paths_t[ind] = roc.threshold_probs(probs[rep][ind], \
threshold, \
sim_args['species_to'], \
ps[rep], 0, \
sim_args['num_sites']-1)
predicted = sim_process.convert_to_blocks(paths_t, predict_args['states'])
stats = roc.get_stats(actual[rep], predicted, sim_args)
all_stats.append(stats)
avg_stats = {}
for key in all_stats[0].keys():
x = []
for rep in range(sim_args['num_reps']):
x.append(all_stats[rep][key])
avg_stats[key] = mystats.mean(x)
roc.write_roc_line(f, threshold, avg_stats, header)
header = False
f.close()
# ======
# read in simulated sequences
# ======
# trees, recomb_sites, seg_sites, positions, seqs
sim = sim_process.read_one_sim(ms_f, sim_args['num_sites'],
sim_args['num_samples'])
# ======
# predict introgressed/non-introgressed tracts
# ======
state_seq, hmm, hmm_init = sim_predict.predict_introgressed(
sim, sim_args, predict_args, train=True, method="viterbi")
state_seq_blocks = sim_process.convert_to_blocks(state_seq,
predict_args['states'])
# ======
# output
# ======
# summary info about HMM (before training)
sim_predict.write_hmm_line(hmm_init, out_init_f, i == 0)
# summary info about HMM (after training)
sim_predict.write_hmm_line(hmm, out_f, i == 0)
# locations of introgression
sim_process.write_introgression_blocks(state_seq_blocks, introgression_f,
i, predict_args['states'])
# x = 20
# thresholds = [float(i) / x for i in range(0, x + 2)]
thresholds = [
0, .00001, .00005, .0001, .0005, .001, .005, .01, .05, .1, .5, .6, .7, .8,
.9, .99, 1, 1.1
]
header = True
for threshold in thresholds:
print('threshold:', threshold)
all_stats = []
for rep in range(sim_args['num_reps']):
paths_t = {}
for ind in probs[rep]:
paths_t[ind] = roc.threshold_probs(probs[rep][ind], threshold,
sim_args['species_to'], ps[rep],
0, sim_args['num_sites'] - 1)
predicted = sim_process.convert_to_blocks(paths_t,
predict_args['states'])
stats = roc.get_stats(actual[rep], predicted, sim_args)
all_stats.append(stats)
avg_stats = {}
for key in all_stats[0].keys():
x = []
for rep in range(sim_args['num_reps']):
x.append(all_stats[rep][key])
avg_stats[key] = mystats.mean(x)
roc.write_roc_line(f, threshold, avg_stats, header)
header = False
f.close()
# ======
stats = sim_stats(sim, args)
# ======
# calculate frequency of ILS (or of possible ILS...)
# ======
concordance_info = calculate_ils(sim, args)
# ======
# find introgressed/non-introgressed tracts
# ======
introgression_stats, actual_state_seq = find_introgressed(sim, args)
actual_state_seq_blocks = sim_process.convert_to_blocks(
actual_state_seq, args['species'])
# ======
# output
# ======
# general summary statistics about simulated sequences
write_output_line(stats, concordance_info, introgression_stats, out_f,
i == 0)
# specific locations of introgression (for comparing predictions
# to)
sim_process.write_introgression_blocks(actual_state_seq_blocks,
introgression_f, i, args['species'])
ms_f.close()
##======
stats = sim_stats(sim, args)
##======
# calculate frequency of ILS (or of possible ILS...)
##======
concordance_info = calculate_ils(sim, args)
##======
# find introgressed/non-introgressed tracts
##======
introgression_stats, actual_state_seq = find_introgressed(sim, args)
actual_state_seq_blocks = sim_process.convert_to_blocks(actual_state_seq, \
args['species'])
##======
# output
##======
# general summary statistics about simulated sequences
write_output_line(stats, concordance_info, introgression_stats, out_f, i==0)
# specific locations of introgression (for comparing predictions
# to)
sim_process.write_introgression_blocks(actual_state_seq_blocks, introgression_f, \
i, args['species'])
ms_f.close()
out_f.close()
# read in simulated sequences
# ======
# trees, recomb_sites, seg_sites, positions, seqs
sim = sim_process.read_one_sim(ms_f, sim_args['num_sites'],
sim_args['num_samples'])
# ======
# predict introgressed/non-introgressed tracts
# ======
state_seq, probs, init, emis, trans, ps = \
sim_predict_phylohmm.predict_introgressed(sim, sim_args,
predict_args, i, gp_dir)
state_seq_blocks = sim_process.convert_to_blocks(state_seq,
sim_args['species'])
# ======
# output
# ======
if write_ps:
predict.write_positions(ps, ps_f, str(i), 'I')
# summary info about HMM
# sim_predict.write_hmm_line(hmm, out_f, i==0)
# locations of introgression
sim_process.write_introgression_blocks(state_seq_blocks, introgression_f,
i, sim_args['species'])