def get_motif_from_tsv(genom, config):
"""Create a motif by reading csv file"""
df = pd.read_csv(config.motif_file, sep='\t')
df = df[(df.TF == config.TF)
& (df.genome_accession == genome.accession_full(genom))]
sites = run_lasagna(df, genom)
return motif.new_motif(sites)
def random_motif(ref, target, config):
"""Given reference motif M having k sites, generate k random subsequences
from target genome and build a motif from randomly generated sites"""
true_motif = get_true_motif(ref, config)
mlen = motif.length(true_motif)
msize = motif.size(true_motif)
sites = [genome.random_seq(target, mlen) for _ in range(msize)]
return motif.new_motif(sites)
def meme_on_pssm_searched(ref, target, config, seeded):
"""Scan the target genome with the reference motif. Pass regions with
putative site to the motif discovery as input sequences"""
meme_settings = get_meme_settings(config, ref, seeded)
true_motif = get_true_motif(ref, config)
nsites = int(2.15 * motif.size(true_motif))
psites = motif.pssm_search_on_regions(true_motif, genome.promoters(target),
nsites)
regions = sequence.merge_overlapping_seqs(
[sequence.expand(site) for site in psites])
for reg in regions:
print reg.start, reg.end, reg.strand
motifs = [
motif.new_motif(sites)
for sites in meme.motif_discovery(regions, meme_settings)
]
return motifs
def network_transfer(ref, target, config, seeded):
"""Given reference regulon, identify the regulon that is orthologous to the
reference and therefore the motif. Extension to the network transfer is to
perform motif discovery seeded."""
meme_settings = get_meme_settings(config, ref, seeded)
true_motif = get_true_motif(ref, config)
true_regulon = motif.regulon(true_motif)
# Get orthologs
orthologs = genome.orthologs(ref, target, config.ortholog_dir)
# Get target regulon (inferred)
inf_regulon = listutils.nub([
gene.operon(orthologs[g], target) for opr in true_regulon for g in opr
if orthologs.get(g, None)
])
promoters = sequence.merge_overlapping_seqs(
[gene.upstream_region(target, opr[0]) for opr in inf_regulon])
# If less than 3 promoters, search for motif in all promoters
if len(promoters) < 3:
promoters = genome.promoters(target)
motifs = [
motif.new_motif(sites)
for sites in meme.motif_discovery(promoters, meme_settings)
]
return motifs
def direct_transfer(ref, target, config):
true_motif = get_true_motif(ref, config)
nsites = int(1.15 * motif.size(true_motif))
sites = motif.pssm_search_on_regions(true_motif, genome.promoters(target),
nsites)
return motif.new_motif(sites)