forked from porchard/2020-sn-muscle
-
Notifications
You must be signed in to change notification settings - Fork 2
/
logistic-regression-multiple-states.nf
149 lines (105 loc) · 4.06 KB
/
logistic-regression-multiple-states.nf
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
#!/usr/bin/env nextflow
ROOT = params.projroot
GENOME = params.genome
TSS = params.tss[GENOME]
CHROM_SIZES = params.chrom_sizes[GENOME]
DOWNSTREAM = params.downstream + '/results'
// Generic data
AUTOSOMAL_REFERENCES = ['hg19': (1..22).collect({it -> 'chr' + it}),
'hg38': (1..22).collect({it -> 'chr' + it}),
'rn5': (1..20).collect({it -> 'chr' + it}),
'rn6': (1..20).collect({it -> 'chr' + it}),
'mm9': (1..19).collect({it -> 'chr' + it}),
'mm10': (1..19).collect({it -> 'chr' + it})
]
get_autosomes = {
genome ->
AUTOSOMAL_REFERENCES[genome]
}
// Name master peaks
// Filter master peaks to TSS-distal
// Get accessibility of each (peak, cluster)
// Lift to human coordinates, if necessary
process name_and_filter_peaks {
executor 'local'
container "${params.containers.general}"
input:
file(master_peaks) from Channel.fromPath("${DOWNSTREAM}/process-by-cluster-round-1/master-peaks/master-peaks.${GENOME}.bed")
output:
file('master-peaks.bed') into accessibility_peaks_in
file('master-peaks.bed') into master_peaks_in
"""
bedtools slop -b 5000 -i $TSS -g $CHROM_SIZES | sort -k1,1 -k2n,2 > tss-proximal.bed
cat $master_peaks | awk '{print(\$1, \$2, \$3, "peak_" NR)}' | perl -pe 's/ /\\t/g' | sort -T . -k1,1 -k2n,2 | bedtools intersect -a stdin -b tss-proximal.bed -v > master-peaks.bed
"""
}
liftover_in = Channel.create()
not_lifted = Channel.create()
master_peaks_in.choice(liftover_in, not_lifted) { a -> GENOME == 'rn6' ? 0 : 1}
process lift_master_peaks {
executor 'local'
container "${params.containers.bnmapper}"
input:
file(master_peaks) from liftover_in
output:
file('master-peaks.lifted-sorted.bed') into lifted
"""
bedtools slop -b 5000 -i ${params.tss['hg19']} -g ${params.chrom_sizes['hg19']} | sort -k1,1 -k2n,2 > tss-proximal.hg19.bed
bnMapper.py -o master-peaks.lifted.bed $master_peaks ${params.chain}
cat master-peaks.lifted.bed | grep -w -e ${get_autosomes('hg19').join(' -e ')}| sort -k1,1 -k2n,2 -T . | bedtools intersect -a stdin -b tss-proximal.hg19.bed -v > master-peaks.lifted-sorted.bed
"""
}
process enhancer_regression_accessibility {
container "${params.containers.general}"
input:
file(peaks) from Channel.fromPath("${DOWNSTREAM}/process-by-cluster-round-1/peaks/*-${GENOME}_peaks.broadPeak.noblacklist").toSortedList()
file(master_peaks) from accessibility_peaks_in
output:
file("accessibility.bed") into enhancer_regression_accessibility_out
"""
cat ${peaks.join(' ')} | sort -T . -k1,1 -k2n,2 | perl -pe 's/\\t(\\d+)-${GENOME}_peak.*/\\tcluster_\$1/' > cluster-peaks.txt
bedtools intersect -a $master_peaks -b cluster-peaks.txt -sorted -wa -wb | cut -f4,8 > accessibility.bed # peakname, cluster it's accessible in (one per line)
"""
}
// Run the enhancer regression
process get_enhancer_posteriors {
memory '10 GB'
maxForks 50
tag "${cell_type} ${chrom}"
container "${params.containers.general}"
input:
file(master_peaks) from lifted.mix(not_lifted)
each file(posterior_file) from Channel.fromPath("${ROOT}/data/roadmap-posteriors/*posterior.txt.gz")
output:
set val(cell_type), file("scores.bed") into enhancer_posteriors_out
script:
m = posterior_file.getName() =~ /(.*)_15_coreMarks_(chr.*)_posterior.txt.gz/
cell_type = m[0][1]
chrom = m[0][2]
"""
fetch-enhancer-posteriors-multiple-states.py $master_peaks $posterior_file E6 E7 E12 > scores.bed
"""
}
process concat_enhancer_posteriors {
tag "${cell_type}"
container "${params.containers.general}"
input:
set val(cell_type), file("scores*.bed") from enhancer_posteriors_out.groupTuple()
output:
file("${cell_type}.bed") into enhancer_posteriors
"""
cat scores*.bed | grep -v '^name' | sort -T . > ${cell_type}.bed
"""
}
process enhancer_regression {
publishDir "${params.results}/enhancer-regression"
memory '20 GB'
input:
file(accessibility) from enhancer_regression_accessibility_out
file(posteriors) from enhancer_posteriors.toSortedList()
output:
file("model_results.txt") into enhancer_regression_out
"""
logistic-regression-on-enhancer-posteriors.py $accessibility ${posteriors.join(' ')} | grep cluster > model_results.txt
"""
}