updating figure notebooks

src/evaluation/figure_notebooks/figure_5/chip_seq_compare/atac_to_chip.txt

@@ -0,0 +1,40 @@
+GM12878	IRF4	ENCSR000BGY
+GM12878	CTCF	ENCSR000DZN
+GM12878	SPI1	ENCSR000BGQ
+GM12878	RUNX	ENCSR000BRI
+GM12878	JDP2	ENCSR000BGT
+GM12878	NFKB1	ENCSR387QUV
+GM12878	IRF4	None
+GM12878	ELK1	ENCSR000DZB
+GM12878	SP1	ENCSR000BHK
+GM12878	NFYA	ENCSR000DNN
+K562	GATA1	ENCFF148JKK
+K562	CTCF	ENCSR000EGM
+K562	SP1	ENCSR991ELG
+K562	NFE2L1	ENCSR632SHZ
+K562	ELK1	ENCFF119SCQ
+K562	NFYB	ENCSR000EGQ
+K562	CEBPG	ENCSR490LWA
+K562	None	None
+K562	NRF1	ENCFF626VDA
+K562	BACH1	ENCFF543FNN
+HEPG2	CTCF	ENCSR000BIE
+HEPG2	HNF4G	ENCSR000BNJ
+HEPG2	FOXM1	ENCSR435ARI
+HEPG2	SP1	ENCSR334KIQ
+HEPG2	FOXO3	ENCSR321OAA
+HEPG2	HNF1B	ENCSR127XTZ
+HEPG2	ATF3	ENCSR000BKE
+HEPG2	NFYB	ENCSR935GZV
+HEPG2	ELK1	ENCSR717QSS
+HEPG2	CEBPG	ENCSR639IIZ
+H1ESC	CTCF	ENCSR000AMF
+H1ESC	POU5F1	ENCSR000BMU
+H1ESC	ZIC3	None
+H1ESC	SP1	ENCSR000BIR
+H1ESC	NANOG	ENCSR000BMT
+H1ESC	TEAD1	ENCSR000BRY
+H1ESC	SOX2	None
+H1ESC	NFY	None
+H1ESC	RFX	ENCSR000ECF
+H1ESC	SIX	ENCSR000BIQ

src/evaluation/figure_notebooks/figure_5/chip_seq_compare/atac_to_chip_1.txt

@@ -0,0 +1,84 @@
+GM12878	IRF4	ENCSR000BGY 
+GM12878	CTCF	ENCSR000DZN
+GM12878	SPI1	ENCSR000BGQ
+GM12878	RUNX3	ENCSR000BRI
+GM12878	BATF.JUNB	ENCSR000BGT
+GM12878	RELB	ENCSR387QUV
+GM12878	IRF4.SPI1	ENCSR000BGY
+GM12878	GABPA	ENCSR331HPA
+GM12878	SP1	ENCSR000BHK
+GM12878	NFYA	ENCSR000DNN
+GM12878	RELB	ENCSR387QUV
+GM12878	BATF	ENCSR000BGT
+GM12878	NRF1	ENCSR000DZO
+GM12878	POU2F2	ENCSR000BGP
+GM12878	ATF3.CREB1	ENCSR000BJY
+GM12878	RELB	ENCSR387QUV
+GM12878	BATF	ENCSR000BGT
+GM12878	ZNF143	ENCSR936XTK
+GM12878	EBF1	ENCSR000BGU
+GM12878	None	None
+GM12878	PAX5	ENCSR000BHJ
+K562	GATA1.TAL1	ENCFF148JKK
+K562	CTCF	ENCSR000EGM
+K562	SP1	ENCSR991ELG
+K562	NFE2L1	ENCSR632SHZ
+K562	ELK1	ENCFF119SCQ
+K562	NFYA	ENCSR163VTS
+K562	CEBPG	ENCSR490LWA
+K562	SPI1	ENCSR000BGW
+K562	NRF1	ENCFF626VDA
+K562	NFE2	ENCSR000FCC
+K562	ATF2	ENCSR014ARU
+K562	NFIX	ENCSR574VJG
+K562	ZNF143	ENCSR427WZJ
+K562	None	None
+K562	YY1	ENCSR000EWF
+K562	STAT5A	ENCSR000BRR
+K562	ZBTB33	ENCSR876GXA
+K562	SP1	ENCSR991ELG
+K562	TFE3	ENCSR953KEY
+K562	ATF2	ENCSR014ARU
+K562	ZNF143	ENCSR427WZJ
+HEPG2	CTCF	ENCSR000BIE,ENCSR607XFI,ENCSR019NPF,ENCSR000DUG,ENCSR000AMA,ENCSR000DLS
+HEPG2	HNF4A	ENCSR469FBY
+HEPG2	FOXA2	ENCSR066EBK
+HEPG2	SP5	ENCSR019NPF
+HEPG2	FOXA1
+HEPG2	HNF1A
+HEPG2	JUND
+HEPG2	NFYB
+HEPG2	GABPA
+HEPG2	CEBPB
+HEPG2	NR5A1
+HEPG2	NR5A1
+HEPG2	FOXA1
+HEPG2	FOXA2
+HEPG2	SOX13
+HEPG2	FOXA1
+HEPG2	NRF1
+HEPG2	TEAD4
+HEPG2	ZNF143
+HEPG2	CEBPB
+HEPG2	ATF4
+H1ESC	CTCF
+H1ESC	POU5F1
+H1ESC	None
+H1ESC	SP1
+H1ESC	POU5F1.NANOG
+H1ESC	TEAD4
+H1ESC	NANOG
+H1ESC	None
+H1ESC	RFX5
+H1ESC	SIX5
+H1ESC	TEAD4
+H1ESC	ZNF143
+H1ESC	JUN
+H1ESC	NRF1
+H1ESC	RFX5
+H1ESC	REST
+H1ESC	CTCF
+H1ESC	None
+H1ESC	GABPA
+H1ESC	None
+H1ESC	JUN

src/evaluation/figure_notebooks/figure_5/chip_seq_compare/compute_imp_scores.py

@@ -0,0 +1,92 @@
+import argparse
+import pandas as pd
+import pyBigWig
+import pybedtools
+import numpy as np
+
+parser=argparse.ArgumentParser(description="")
+parser.add_argument("-ac","--atac_counts", required=True,  help="")
+parser.add_argument("-ap","--atac_profile", required=True,  help="")
+parser.add_argument("-dc","--dnase_counts", required=True,  help="")
+parser.add_argument("-dp","--dnase_profile", required=True,  help="")
+parser.add_argument("-cb","--chip_bed", required=True,  help="")
+parser.add_argument("-cc","--chip_contrib", required=True,  help="")
+parser.add_argument("-tb","--tobias_bed", required=True,  help="")
+parser.add_argument("-o","--outf", required=True,  help="")
+parser.add_argument("-bwo","--observed", required=True,  help="")
+parser.add_argument("-dbwo","--dnobserved", required=True,  help="")
+args = parser.parse_args()
+
+
+
+if __name__=="__main__":
+
+	a = pybedtools.example_bedtool(args.tobias_bed)
+	#b = pybedtools.example_bedtool(args.chip_bed)
+	bdf = pd.read_csv(args.chip_bed, sep='\t', header=None)
+	bdf[1] = bdf[1] + bdf[9] - 500
+	bdf[2] = bdf[1] + 1000
+	bdf[bdf[1] < 0][1] = 0
+
+	b = pybedtools.BedTool.from_dataframe(bdf)
+	a_and_b = a.intersect(b, c=True, f=1.0)
+	result = a_and_b.to_dataframe()
+	print(result.head())
+	result['label'] = result["blockSizes"] > 0
+
+	tobia_bed = pd.read_csv(args.tobias_bed, sep="\t", header=None)
+	bw_counts = pyBigWig.open(args.atac_counts)
+	bw_profile = pyBigWig.open(args.atac_profile)
+	dn_bw_counts = pyBigWig.open(args.dnase_counts)
+	dn_bw_profile = pyBigWig.open(args.dnase_profile)
+	bw_chip = pyBigWig.open(args.chip_contrib)
+	bw_obs = pyBigWig.open(args.observed)
+	dn_bw_obs = pyBigWig.open(args.dnobserved)
+
+
+	chip_seq_vals = []
+	atac_counts = []
+	atac_profiles = []
+	obs_vals = []
+	dnase_counts = []
+	dnase_profiles = []
+	dnase_obs_vals = []
+
+	tobia_bed = tobia_bed[result['label']].reset_index(drop=True)
+
+	for i,r in tobia_bed.iterrows():
+		val1 = np.sum(np.nan_to_num(bw_counts.values(r[0],r[1],r[2])))
+		atac_counts.append(val1)
+
+		val1 = np.sum(np.nan_to_num(bw_profile.values(r[0],r[1],r[2])))
+		atac_profiles.append(val1)
+
+		val1 = np.sum(np.nan_to_num(dn_bw_counts.values(r[0],r[1],r[2])))
+		dnase_counts.append(val1)
+
+		val1 = np.sum(np.nan_to_num(dn_bw_profile.values(r[0],r[1],r[2])))
+		dnase_profiles.append(val1)
+
+		val1 = np.sum(np.nan_to_num(bw_chip.values(r[0],r[1],r[2])))
+		chip_seq_vals.append(val1)
+
+		mid=int((r[1]+r[2])/2)
+		val1 = np.sum(np.nan_to_num(bw_obs.values(r[0],mid-150,mid+150)))
+		obs_vals.append(val1)
+
+
+		mid=int((r[1]+r[2])/2)
+		val1 = np.sum(np.nan_to_num(dn_bw_obs.values(r[0],mid-150,mid+150)))
+		dnase_obs_vals.append(val1)
+
+	tobia_bed["counts_contrib"] = atac_counts
+	tobia_bed["profiles_contrib"] = atac_profiles
+	tobia_bed["dnase_counts_contrib"] = dnase_counts
+	tobia_bed["dnase_profiles_contrib"] = dnase_profiles
+	tobia_bed["chip"] = chip_seq_vals
+	tobia_bed["observed"] = obs_vals
+	tobia_bed["dnase_observed"] = dnase_obs_vals
+	#tobia_bed["label"] = result['label']
+
+
+	tobia_bed.to_csv(args.outf, sep='\t', header=False, index=False)