def test_Fst__windowed(sample_size, n_cohorts, chunks):
ts = simulate_ts(sample_size, length=200)
ds = ts_to_dataset(ts, chunks) # type: ignore[no-untyped-call]
ds, subsets = add_cohorts(ds, ts,
n_cohorts) # type: ignore[no-untyped-call]
ds = window(ds, size=25)
fst_ds = Fst(ds, estimator="Nei")
fst = fst_ds["stat_Fst"].values
# Calculate Fst using tskit windows
# Find the variant positions so we can have windows with a fixed number of variants
positions = ts.tables.sites.position
windows = np.concatenate(([0], positions[::25][1:], [ts.sequence_length]))
n_windows = len(windows) - 1
ts_fst = np.full([n_windows, n_cohorts, n_cohorts], np.nan)
for i, j in itertools.combinations(range(n_cohorts), 2):
ts_fst[:, i, j] = ts.Fst([subsets[i], subsets[j]],
windows=windows,
span_normalise=False)
ts_fst[:, j, i] = ts_fst[:, i, j]
# We can values close to zero, and the default value of atol isn't
# appropriate for this.
atol = 1e-8
np.testing.assert_allclose(fst, ts_fst, atol=atol)
# scikit-allel
fst_ds = Fst(ds, estimator="Hudson")
for i, j in itertools.combinations(range(n_cohorts), 2):
fst = fst_ds["stat_Fst"].sel(cohorts_0=f"co_{i}",
cohorts_1=f"co_{j}").values
ac_i = fst_ds.cohort_allele_count.values[:, i, :]
ac_j = fst_ds.cohort_allele_count.values[:, j, :]
ska_fst = allel.moving_hudson_fst(ac_i, ac_j, size=25)
np.testing.assert_allclose(
fst[:-1], ska_fst,
atol=atol) # scikit-allel has final window missing
chrom=chrom,
samples=metadata,
numbers=numbers,
ploidy=ploidy,
qualflt=qualflt,
missingfltprop=missingprop)
#### Fst in windows ####
for sus, res in comparisons:
name = sus + "_" + res
cohortText = f"{sus} v {res}"
print(f"Calculating Fst values in sliding windows for {name}\n")
for wname, size, step in zip(windownames, windowsizes, windowsteps):
FstArray = allel.moving_hudson_fst(acsubpops[sus],
acsubpops[res],
size=size,
step=step)
midpoint = allel.moving_statistic(pos,
np.median,
size=size,
step=step)
cohortNoSpaceText = name + "." + wname
rnaseqpop.plotWindowed(
statName="Fst",
cohortText=cohortText,
cohortNoSpaceText=cohortNoSpaceText,
values=FstArray,
midpoints=midpoint,
colour='dodgerblue',
prefix="results/variantAnalysis/selection/fst",
def main(args):
## Step 0: get null model for SNP calling
null_loc = os.path.dirname(
__file__) + '/helper_files/combined_null1000000.txt'
null_model = generate_snp_model(null_loc)
P2C = {'A': 0, 'C': 1, 'T': 2, 'G': 3}
C2P = {0: 'A', 1: 'C', 2: 'T', 3: 'G'}
## Step 1: build new counts table from all objects
s_final = SNPprofile()
s_final.filename = args.output
i = 0
counts_per_block = {}
s1 = SNPprofile()
print("loading " + args.input[0])
s1.load(args.input[0])
s_final.scaffold_list = s1.scaffold_list
s_final.counts_table = copy.deepcopy(s1.counts_table)
s2 = SNPprofile()
print("loading " + args.input[1])
s2.load(args.input[1])
for scaf in s2.scaffold_list:
if scaf not in s_final.scaffold_list:
sys.exit(
"Error: scaffold " + scaf + " in " + fn +
" not found in initial file. Your inStrain objects were probably not run on the same FASTA."
)
scaf_counter = 0
for scaf in s2.counts_table:
s_final.counts_table[scaf_counter] += scaf
scaf_counter += 1
i += 1
# Step 2: call all SNPs for new object
allele_counts_total = {}
allele_counts1 = {}
allele_counts2 = {}
snp_table = defaultdict(list)
scaf_counter = 0
for scaf in tqdm(s_final.counts_table, desc='Calling new SNVs...'):
pos_counter = 0
for counts in scaf:
snp = call_snv_site(counts,
min_cov=5,
min_freq=0.05,
model=null_model)
if snp: # means that there was coverage at this position
if snp != -1: # means this is a SNP
# calculate varBase
snp, varbase = major_minor_allele(counts)
snp_table['scaffold'].append(
s_final.scaffold_list[scaf_counter])
snp_table['position'].append(pos_counter)
snp_table['varBase'].append(snp)
snp_table['conBase'].append(varbase)
allele_counts_total[s_final.scaffold_list[scaf_counter] +
":" + str(pos_counter)] = (
s_final.counts_table[scaf_counter]
[pos_counter])
allele_counts1[s_final.scaffold_list[scaf_counter] + ":" +
str(pos_counter)] = (
s1.counts_table[scaf_counter]
[pos_counter])
allele_counts2[s_final.scaffold_list[scaf_counter] + ":" +
str(pos_counter)] = (
s2.counts_table[scaf_counter]
[pos_counter])
pos_counter += 1 # 0 based positions!!
scaf_counter += 1
# Step 3: Save new FST_SNP table to disk.
SNPTable = pd.DataFrame(snp_table)
FstTable = defaultdict(list)
for gene in tqdm(create_gene_index(args.gene_file),
desc="calculating fst"):
snps = SNPTable[(SNPTable.scaffold == gene['scaf'])
& (SNPTable.position >= gene['start']) &
(SNPTable.position <= gene['end'])]
snp_list = []
for index, row in snps.iterrows():
snp_list.append(row['scaffold'] + ":" + str(row['position']))
# only continue if there are at least 3 snps in this gene
if len(snp_list) >= 3:
allele_counts_1 = []
allele_counts_2 = []
for snp in snp_list:
allele_counts_1.append(allele_counts1[snp])
allele_counts_2.append(allele_counts2[snp])
allel1 = allel.AlleleCountsArray(allele_counts_1)
allel2 = allel.AlleleCountsArray(allele_counts_2)
fst_h = allel.moving_hudson_fst(
allel1, allel2,
size=len(snp_list))[0] #allel.moving_hudson_fst(a1,a2, size=3)
nd_1 = np.sum(allel.mean_pairwise_difference(allel1)) / (
1 + gene['end'] - gene['start'])
nd_2 = np.sum(allel.mean_pairwise_difference(allel2)) / (
1 + gene['end'] - gene['start'])
FstTable['gene'].append(gene['name'])
FstTable['snp_num'].append(len(snp_list))
FstTable['fst'].append(fst_h)
FstTable['pi_1'].append(nd_1)
FstTable['pi_2'].append(nd_2)
FstTable['cov_1'].append(np.mean(np.sum(allele_counts_1, axis=1)))
FstTable['cov_2'].append(np.mean(np.sum(allele_counts_2, axis=1)))
FstTable = pd.DataFrame(FstTable)
print(np.mean(FstTable['fst']))
FstTable.to_csv(args.output + '.Fst.tsv', index=False, sep='\t')
