Coder Social home page Coder Social logo

circscan's Introduction

CIRCScan

Tools for predicting circRNAs expression with EPIGENETIC features by Machine Learning


Introduction

Circular RNAs ( circRNAs ) are an abundant class of noncoding RNAs with the widespread, cell/tissue specific pattern.
This tool, CIRCScan, is used for predicting circRNAs expression in a cell/tissue specific manner by machine learning based on epigenetic features.


Citation

Chen JB*, Dong SS*, Yao S, Duan YY, Hu WX, Chen H, Wang NN, Chen XF, Hao RH, Thynn HN, Guo MR, Zhang YJ, Rong Y, Chen YX, Zhou FL, Guo Y#, Yang TL#. Modeling circRNA expression pattern with integrated sequence and epigenetic features demonstrates the potential involvement of H3K79me2 in circRNA expression. Bioinformatics, 2020.


License

This software is distributed under the terms of GPL 3.0


Source

https://github.com/johnlcd/CIRCScan


Contact

Author

Jia-Bin Chen, Shan-Shan Dong, Yan Guo, Tie-Lin Yang
Key Laboratory of Biomedical Information Engineering of Ministry of Education, School of Life Science and Technology, Xi'an Jiaotong University, Xi'an, Shaanxi Province, 710049, P. R. China
📧 [email protected]


Maintainer

Jia-Bin Chen
You can contact 📧 [email protected] when you have any questions, suggestions, comments, etc.
Please describe in details, and attach your command line and log messages if possible.


Requiremnets

  • bedtools ( v2.25.0 )
  • Python ( recommended: python2.7 )
  • R ( >= 3.2.4 )
    • R packages: caret (6.0-73), ggplot2 (2.2.0), doParallel (3.2.4), ROCR (1.0-7), etc. ( Dependent packages for different models )

Check the log file ".out" to validate which package is required if get an error info


Building CIRCscan

CMD:

	git clone https://github.com/johnlcd/CIRCScan.git

Directory catalog

  • bin
    • anno
      • alu_anno_IP.py
      • anno_bedpe.py
      • anno_intron.py
      • bt_intersect_alu_intron.sh
      • bt_overlap
      • comb_pair_anno.py
    • anno_pair
    • circscan
    • merge_feature
    • model
      • Circ_pred.R
      • Eval_test_cor.R
      • Eval_test_perf.R
      • feature_selection.R
      • make_set.py
      • Model_train_FS_exp.R
      • Model_train_FS_exp_rep.R
      • Model_train_FS_exp_rmo.R
      • Model_train_FS_exp_rmo_rep.R
      • Model_train_pred_exp.R
      • Model_train.R
      • Model_train_rep.R
    • prepare_train_set
  • data
    • Alu
      • alu_hg19.bed
      • pair_anno_alu
    • anno
      • GM12878_anno_comb.gz
      • H1-hESC_anno_comb.gz
      • HeLa-S3_anno_comb.gz
      • HepG2_anno_comb.gz
      • K562_anno_comb.gz
      • NHEK_anno_comb.gz
    • DNaseI
      • GM12878_dnase.bed
      • H1-hESC_dnase.bed
      • HeLa-S3_dnase.bed
      • HepG2_dnase.bed
      • HUVEC_dnase.bed
      • K562_dnase.bed
      • NHEK_dnase.bed
    • histone
      • A549_his.bed
      • GM12878_his.bed
      • H1-hESC_his.bed
      • HeLa-S3_his.bed
      • HepG2_his.bed
      • HUVEC_his.bed
      • K562_his.bed
      • NHEK_his.bed
    • known_circ
      • A549_circ_overlap.bed
      • GM12878_circ_overlap.bed
      • H1-hESC_circ_overlap.bed
      • HeLa-S3_circ_overlap.bed
      • HepG2_circ_overlap.bed
      • HMEC_circ_overlap.bed
      • HOB_circ_overlap.bed
      • HSMM_circ_overlap.bed
      • HUVEC_circ_overlap.bed
      • K562_circ_overlap.bed
      • NHA_circ_overlap.bed
      • NHDF_circ_overlap.bed
      • NHEK_circ_overlap.bed
      • NHLF_circ_overlap.bed
    • pred_true_bycell
      • GM12878_pred_true.bed.gz
      • H1-hESC_pred_true.bed.gz
      • HeLa-S3_pred_true.bed.gz
      • HepG2_pred_true.bed.gz
      • K562_pred_true.bed.gz
      • NHEK_pred_true.bed.gz
    • raw_data
      • DNaseI.txt.gz
      • Histone_part1.txt.gz
      • Histone_part2.txt.gz
      • select_cell.list
      • select_his.list
  • info
    • models_ALL.txt
    • models_classification.list
    • models_test.list
    • models_test.txt
  • README.md
  • sample
    • anno
    • feature
    • model

Runing preparation

  • Set environment variables

CMD:

	export PKG_DIR=/path/to/CIRCScan
	export PATH=$PKG_DIR/bin:PKG_DIR/bin/anno:$PKG_DIR/bin/model:$PATH

  • Unzip data files

CMD:

	cd $PKG_DIR/data
	tar -zxvf intron_pairs_data.tgz
	cd $PKG_DIR/data/raw_data
	gunzip *.gz
	cat Histone_part1.txt Histone_part2.txt > Histone.txt

Work flow

CIRCScan Pipeline

CIRCScan pipeline

1. Data preparation and feature generation

Epigenetic data including DNaseI HS, Histone modification, downloaded from ENCODE http://hgdownload.cse.ucsc.edu/goldenPath/hg19/encodeDCC/

  • Process the raw broadPeak/narrowPeak file of ENCODE

CMD:

For each histone mark: (e.g.: H3K36me3)

	zcat wgEncodeBroadHistoneK562H3k36me3StdPk.broadPeak.gz | awk -v OFS='\t' '{print "K562","None","H3K36me3",$1,$2,$3}' > K562_H3K36me3.txt

then merged all marks of selected cell lines:

	#rm -f Histon.txt
	cat <(echo "cell treatment antibody chr start end" | sed 's/ /\t/g') *.txt > Histon.txt (Put these '.txt' file in one directory)

For DHS:

	zcat wgEncodeAwgDnaseUwdukeK562UniPk.narrowPeak.gz | awk -v OFS='\t' '{print "K562","None",$1,$2,$3}' > K562_DNaseI.txt

then merged all cell line:

	#rm -f DNaseI.txt
	cat <(echo "cell treatment chr start end" | sed 's/ /\t/g') *.txt > DNaseI.txt

  • Extract features data from .txt file, transform into BED fromate

CMD:

	grep K562 Histon.txt | grep -f $PKG_DIR/data/raw_data/select_his.list | awk -v OFS='\t' '{print $4,$5,$6,$3}' > K562_his.bed

	grep K562 DNaseI.txt | awk -v OFS='\t' '{print $4,$5,$6,"DNaseI_HS"}' > K562_dnase.bed

"Histon.txt"

cell	treatment	antibody	chr	start	end  
GM12878	None	CTCF	chr22	16846634	16869580  
GM12878	None	CTCF	chr22	16850639	16850924  
GM12878	None	CTCF	chr22	16851700	16851834  
GM12878	None	CTCF	chr22	16852344	16852458  
GM12878	None	CTCF	chr22	16853076	16853192  
GM12878	None	CTCF	chr22	16853755	16853871  
GM12878	None	CTCF	chr22	16854517	16854638  
GM12878	None	CTCF	chr22	16857119	16857231  
GM12878	None	CTCF	chr22	16857764	16857871  
...

"DNaseI.txt"

lab	cell	treatment	chr	start	end
Duke	8988T	None	chr1	564665	564815
Duke	8988T	None	chr1	565025	565175
Duke	8988T	None	chr1	565865	566015
Duke	8988T	None	chr1	714005	714155
Duke	8988T	None	chr1	762785	762935
Duke	8988T	None	chr1	766705	766855
Duke	8988T	None	chr1	767945	768095
Duke	8988T	None	chr1	794145	794295
Duke	8988T	None	chr1	795945	796095
...

"K562_his.bed"

chr1	10140	10374	H3K9me3
chr1	118494	118714	H3K9ac
chr1	118556	118713	H3K4me3
chr1	137502	140080	H3K9ac
chr1	138030	140084	H3K4me2
chr1	138411	138738	H3K4me3
chr1	138424	138651	H3K27ac
chr1	138426	138651	H3K79me2
chr1	138934	139174	H3K4me3
chr1	138938	139177	CTCF
...

"K562_dnase.bed"

chr1	115600	115750	DNaseI_HS
chr1	136280	136430	DNaseI_HS
chr1	138960	139110	DNaseI_HS
chr1	235040	235190	DNaseI_HS
chr1	235600	235750	DNaseI_HS
chr1	237640	237790	DNaseI_HS
chr1	521460	521610	DNaseI_HS
chr1	564480	564630	DNaseI_HS
chr1	565280	565430	DNaseI_HS
chr1	565860	566010	DNaseI_HS
...

  • Feature generation and annotation:

1. Histone modifications, DNaseI HS ... ( Feature types of "bed" format )

Make feature list, and overlap intron with feature, annotate intron by features, then combine intron annotation to pair ( "anno_pair" )

CMD:

	anno_pair -t <cell_type> -f <feature (his, dnase ...)> [ --is (Ignor strands) ] --bed <feature.bed>

	e.g.:
	anno_pair -t K562 -f his --is --bed K562_his.bed

	anno_pair -t K562 -f dnase --is --bed K562_dnase.bed

Generate 4 files:

"K562_his.list"

CTCF  
EZH2_(39875)  
H2A.Z  
H3K27ac  
H3K27me3  
H3K36me3  
H3K4me1  
H3K4me2  
H3K4me3  
H3K79me2  
H3K9ac  
H3K9me3  
H4K20me1

"overlap_K562_his"

chr1	709660	713663	LOC100288069-1-1	1.00	-	chr1	712769	712874	H3K79me2	105
chr1	709660	713663	LOC100288069-1-1	1.00	-	chr1	713056	713748	H3K4me3	607
chr1	709660	713663	LOC100288069-1-1	1.00	-	chr1	713188	713524	H3K27ac	336
chr1	709660	713663	LOC100288069-1-1	1.00	-	chr1	713195	713556	H3K4me2	361
chr1	709660	713663	LOC100288069-1-1	1.00	-	chr1	713199	713548	H3K79me2	349
chr1	709660	713663	LOC100288069-1-1	1.00	-	chr1	713205	713560	H3K9ac	355
chr1	709660	713663	LOC100288069-1-1	1.00	-	chr1	713575	713751	H3K27ac	88
chr1	709660	713663	LOC100288069-1-1	1.00	-	chr1	713578	713747	H3K4me2	85
chr1	709660	713663	LOC100288069-1-1	1.00	-	chr1	713578	713752	H3K79me2	85
chr1	709660	713663	LOC100288069-1-1	1.00	-	chr1	713579	713759	H3K9ac	84
...

"intron_anno_K562_his"

Chr	Start	End	INTRON	CTCF	EZH2_(39875)	H2A.Z	H3K27ac	H3K27me3	H3K36me3	H3K4me1	H3K4me2	H3K4me3	H3K79me2 H3K9ac	H3K9me3	H4K20me1
chr1	12227	12612	DDX11L1-1-1	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000
chr1	12721	13220	DDX11L1-1-2	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000
chr1	15038	15795	WASH7P-1-9	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000
chr1	15947	16606	WASH7P-1-8	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000
chr1	18366	24737	WASH7P-1-2	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000
chr1	24891	29320	WASH7P-1-1	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000
chr1	700627	701708	LOC100288069-1-6	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000
chr1	701767	703927	LOC100288069-1-5	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000
chr1	703993	704876	LOC100288069-1-4	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000
...

"pair_anno_K562_his"

Intron_pair	CTCF	EZH2_(39875)	H2A.Z	H3K27ac	H3K27me3	H3K36me3	H3K4me1	H3K4me2	H3K4me3	H3K79me2	H3K9ac	H3K9me3 H4K20me1
A1BG-1-4_3	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000
A1BG-1-5_3	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000
A1BG-1-5_4	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000
A1BG-1-6_3	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000
A1BG-1-6_4	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000
A1BG-1-6_5	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000
A1BG-1-7_3	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000
A1BG-1-7_4	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000
A1BG-1-7_5	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000
...

"K562_dnase.list"

DNaseI_HS

"overlap_K562_dnase"

chr1	705092	708355	LOC100288069-1-3	1.00	-	chr1	706265	706415	DNaseI_HS	150
chr1	764484	776579	LINC01128-3-2	1.00	+	chr1	767140	767290	DNaseI_HS	150
chr1	764484	783033	LINC01128-2-2	1.00	+	chr1	767140	767290	DNaseI_HS	150
chr1	764484	783033	LINC01128-2-2	1.00	+	chr1	778220	778370	DNaseI_HS	150
chr1	764484	787306	LINC01128-1-2	1.00	+	chr1	767140	767290	DNaseI_HS	150
chr1	764484	787306	LINC01128-1-2	1.00	+	chr1	778220	778370	DNaseI_HS	150
chr1	764484	787306	LINC01128-1-2	1.00	+	chr1	785040	785190	DNaseI_HS	150
chr1	764484	787306	LINC01128-4-2	1.00	+	chr1	767140	767290	DNaseI_HS	150
chr1	764484	787306	LINC01128-4-2	1.00	+	chr1	778220	778370	DNaseI_HS	150
chr1	764484	787306	LINC01128-4-2	1.00	+	chr1	785040	785190	DNaseI_HS	150
...

"intron_anno_K562_dnase"

Chr	Start	End	INTRON	DNaseI_HS
chr1	12227	12612	DDX11L1-1-1	0.000
chr1	12721	13220	DDX11L1-1-2	0.000
chr1	15038	15795	WASH7P-1-9	0.000
chr1	15947	16606	WASH7P-1-8	0.000
chr1	18366	24737	WASH7P-1-2	0.000
chr1	24891	29320	WASH7P-1-1	0.000
chr1	700627	701708	LOC100288069-1-6	0.000
chr1	701767	703927	LOC100288069-1-5	0.000
chr1	703993	704876	LOC100288069-1-4	0.000
...

"pair_anno_K562_dnase"

Intron_pair	DNaseI_HS
A1BG-1-4_3	0.000
A1BG-1-5_3	0.000
A1BG-1-5_4	0.000
A1BG-1-6_3	0.000
A1BG-1-6_4	0.000
A1BG-1-6_5	0.000
A1BG-1-7_3	0.000
A1BG-1-7_4	0.000
A1BG-1-7_5	0.000
...

  • Merge all features

CMD:

	merge_feature -t <cell_type>

	e.g.:
	merge_feature -t K562

Generate "K562_anno_comb", e.g.:

Chr	Start	End	Intron_pair	Alu	DNaseI_HS	CTCF	EZH2_(39875)	H2A.Z	H3K27ac	H3K27me3	H3K36me3	H3K4me1	H3K4me2	H3K4me3	H3K79me2	H3K9ac	H3K9me3	H4K20me1
chr19	58863648	58863921	A1BG-1-4_3	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000
chr19	58862756	58863921	A1BG-1-5_3	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000
chr19	58862756	58863053	A1BG-1-5_4	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000
chr19	58861735	58863921	A1BG-1-6_3	-1.985	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000
chr19	58861735	58863053	A1BG-1-6_4	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000
chr19	58861735	58862017	A1BG-1-6_5	-1.408	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000
chr19	58858718	58863921	A1BG-1-7_3	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000
chr19	58858718	58863053	A1BG-1-7_4	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000
chr19	58858718	58862017	A1BG-1-7_5	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000
...

Newly Updated:

Annotated intron (pair) using raw signal (reads count) from ENCODE "bigWig" file. ("data/anno" directory)

  • Data set preparation for model training, testing, validation and prediction

CMD:

	prepare_train_set -t <cell_type> --circ <known_circ.bed> -R <ratio of negative VS positive> [ --sl < list of intron pair length ( sum of 2 introns ) to do stratified random sampling (space seperated 3 number, defult: 20000 30000 40000 ) > ]
	
	OR:
	
	prepare_train_set -t <cell_type> --circ <known_circ.bed (with expression (SRPBM) of 6 column)> ] --exp ( prepare data sets for expression prediction )

	e.g.:
	prepare_train_set -t K562 --circ K562_circ_overlap.bed -R 1 --sl 30000 50000 70000
	prepare_train_set -t K562 --circ K562_circ_overlap.bed --exp

Generate multiple files: "K562_train", "K562_pred", "K562_circ_intron_pair", "K562_IP_part1", "K562_IP_part2", "K562_IP_part3", "K562_IP_part4"

OR:

"K562_exptrain", "K562_exp_pred"

"K562_train", "K562_pred" used for modeling circRNAs expression status
"K562_exp_train", "K562_exp_pred" used for modeling circRNAs expression levels

"K562_train"

Chr	Start	End	Intron_pair	Alu	DNaseI_HS	CTCF	EZH2_(39875)	H2A.Z	H3K27ac	H3K27me3	H3K36me3	H3K4me1	H3K4me2	H3K4me3	H3K79me2	H3K9ac	H3K9me3	H4K20me1	Type
chr19	58861735	58863053	A1BG-1-6_4	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	F
chr10	52610424	52619745	A1CF-2-3_1	-0.464	0.000	0.000	0.335	0.062	0.000	0.744	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	F
chr10	52595833	52619745	A1CF-2-6_1	-0.850	0.000	0.000	0.482	0.000	0.000	0.716	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	F
chr10	52619601	52619745	A1CF-6-4_3	-0.148	0.000	0.000	0.000	0.000	0.000	0.300	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	F
chr12	9258831	9265132	A2M-1-10_2	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	F
chr12	9256834	9266139	A2M-1-11_1	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	F
chr12	9246060	9262930	A2M-1-18_4	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	F
chr12	9242497	9246175	A2M-1-21_17	0.000	0.000	0.000	0.000	0.000	0.000	1.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	F
chr12	9231839	9247680	A2M-1-25_16	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	F
...

"K562_pred"

Chr	Start	End	Intron_pair	Alu	DNaseI_HS	CTCF	EZH2_(39875)	H2A.Z	H3K27ac	H3K27me3	H3K36me3	H3K4me1	H3K4me2	H3K4me3	H3K79me2	H3K9ac	H3K9me3	H4K20me1	Type
chr19	58863648	58863921	A1BG-1-4_3	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	P
chr19	58862756	58863921	A1BG-1-5_3	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	P
chr19	58862756	58863053	A1BG-1-5_4	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	P
chr19	58861735	58863921	A1BG-1-6_3	-1.985	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	P
chr19	58861735	58862017	A1BG-1-6_5	-1.408	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	P
chr19	58858718	58863921	A1BG-1-7_3	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	P
chr19	58858718	58863053	A1BG-1-7_4	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	P
chr19	58858718	58862017	A1BG-1-7_5	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	P
chr19	58858718	58859006	A1BG-1-7_6	-2.130	0.000	0.000	0.000	0.519	0.000	0.000	0.768	0.000	0.729	0.533	0.000	0.279	0.000	0.000	P
...

"K562_exp_train"

Chr	Start	End	Intron_pair	Alu	DNaseI_HS	CTCF	EZH2_(39875)	H2A.Z	H3K27ac	H3K27me3	H3K36me3	H3K4me1	H3K4me2	H3K4me3	H3K79me2	H3K9ac	H3K9me3	H4K20me1	SRPBM
chr12	125558421	125576069	AACS-1-1_5	0.105	0.015	0.032	1.000	0.041	0.000	0.227	1.000	0.111	0.053	0.000	0.0000.000	0.000	1.000	0.00936076724592655
chr5	178199429	178203277	AACSP1-1-8_4	-6.708	0.000	0.000	0.000	0.000	0.000	0.000	0.454	0.000	0.000	0.000	0.0000.000	0.000	0.000	0.0561646034755593
chr5	178199429	178203277	AACSP1-2-8_4	-6.711	0.000	0.000	0.000	0.000	0.000	0.000	0.454	0.000	0.000	0.000	0.0000.000	0.000	0.000	0.0561646034755593
chr9	99413671	99413994	AAED1-1-4_2	0.000	0.000	0.000	1.000	0.000	0.000	0.000	1.000	0.330	0.000	0.000	0.0000.000	0.221	0.000	0.0280823017377796
chr15	67528316	67529158	AAGAB-1-4_1	-1.888	0.000	0.000	1.000	0.000	0.000	0.000	1.000	0.000	0.000	0.000	1.0000.000	0.996	0.749	0.0374430689837062
chr15	67524151	67529158	AAGAB-1-5_1	0.606	0.000	0.000	1.000	0.000	0.000	0.000	1.000	0.154	0.000	0.000	1.0000.242	0.938	0.492	0.636532172723005
chr15	67500899	67501882	AAGAB-1-7_5	-1.614	0.000	0.000	1.000	0.000	0.000	0.000	1.000	0.168	0.000	0.000	0.9690.562	0.941	0.657	0.0187215344918531
chr15	67528316	67529158	AAGAB-2-4_1	-1.851	0.000	0.000	1.000	0.000	0.000	0.000	1.000	0.000	0.000	0.000	1.0000.000	0.996	0.738	0.0374430689837062
chr15	67524151	67529158	AAGAB-2-5_1	0.594	0.000	0.000	1.000	0.000	0.000	0.000	1.000	0.153	0.000	0.000	1.0000.262	0.938	0.485	0.636532172723005
...

"K562_exp_pred"

Chr	Start	End	Intron_pair	Alu	DNaseI_HS	CTCF	EZH2_(39875)	H2A.Z	H3K27ac	H3K27me3	H3K36me3	H3K4me1	H3K4me2	H3K4me3	H3K79me2	H3K9ac	H3K9me3	H4K20me1	SRPBM
chr19	58863648	58863921	A1BG-1-4_3	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.0000.000	0.000	0.000	EP
chr19	58862756	58863921	A1BG-1-5_3	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.0000.000	0.000	0.000	EP
chr19	58862756	58863053	A1BG-1-5_4	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.0000.000	0.000	0.000	EP
chr19	58861735	58863921	A1BG-1-6_3	-1.985	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.0000.000	0.000	0.000	EP
chr19	58861735	58863053	A1BG-1-6_4	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.0000.000	0.000	0.000	EP
chr19	58861735	58862017	A1BG-1-6_5	-1.408	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.0000.000	0.000	0.000	EP
chr19	58858718	58863921	A1BG-1-7_3	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.0000.000	0.000	0.000	EP
chr19	58858718	58863053	A1BG-1-7_4	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.0000.000	0.000	0.000	EP
chr19	58858718	58862017	A1BG-1-7_5	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.0000.000	0.000	0.000	EP
...

NOTE:

known_circ_overlap.bed: Known circRNAs from RNA-seq data by overlapping the detection of multiple tools


2. Model traing and prediction

  • Complete process for predicting circRNAs expression status

a). Model training ( Used for obtaining rank of importance )

CMD:

	circscan --train -t <cell_type> -m <model> -s/-r <seed>/<run (1-5)> -n <cores>
	# "-n": used for models training by parellel
	# "-s": used for random sample training set ("seed" mode, multiple traning and reproducibility )
	# "-r": used for 5 independent CV run ("run" mode)

	e.g.:
	circscan --train -t K562 -m rf -s 111 -n 8
	or:
	circscan --train -t K562 -m rf -r 1 -n 8

Generate models and R data file "K562_rf_train.RData", log file "K562_rf_train.out" with model evaluation

b). Feature selection

CMD:

	circscan --fs -t <cell_type> -m <model> -n <cores> -l <all/feature_number_list> < --auc / --f1 (referenece index) > [ -pt (type of prediction) raw/prob (probabilities, default) ]	
	# "-n": used for models training by parellel
	# "-l": list of feature number for feature selection. If value is "all", then run feature selection with feature number from 1 to all, if is a list of feature number ( comma separsted ), for example: 1,2,3,4,5,10,15, then run feature selection with feature number you provide
	# "--auc / --f1": referenece index to evaluate model performance
	# "--pt": type of prediction, defult is 'prob' (probabilities), 'raw' is used for models without probabilities

	e.g.:
	circscan --fs -t K562 -m rf -n 8 -l all --auc

Generate R data file "K562_rf_FS.RData" of feature selection and log file "K562_rf_FS.out", and "K562_rf_perf_test.txt" with results of feature selection ( Feature number with highest F1 score )

NOTE:

Feature selection is required to generate and select the best model for circRNAs prediction.

c). Model performance evaluation (testing data)

CMD:

	circscan --eval -t <cell_type> -m <model> -n <cores>
	# "-n": used for models training by parellel

	e.g.:
	circscan --eval -t K562 -m rf -n 8

Generate R data file "K562_rf_test_perf_eval.RData" of model performance evaluation and log file "eval_K562_rf_perf.out", and "K562_rf_eval_test_perf" with results of model performance in testing data

d). CircRNAs prediction and annotation

CMD:

	circscan --pred -t <cell_type> -m <model> -n <cores>
	# "-n": used for models training by parellel

	e.g.:
	circscan --pred -t K562 -m rf -n 8

Generate predicted anaotated circRNAs file "K562_rf_pred_true.bed"


Mode 2). Model training, feature selection, and validation (known circRNAs) (Optional)

CMD:

	circscan --exp -t <cell_type> -m <model> -n <cores> -sf < all/select_fea_list (comma separated)> -l <known circRNA_intron_FIP_list>	
	# "-n": used for models training by parellel
	# "-sf": list of selected feature list (comma separated, default: all)
	# "-l": circBase circRNAs FIP list file

	e.g.:
	circscan --exp -t K562 -m rf -n 8 -sf all/Alu,H3K36me3,... -l GM12878_circ_FIP.list



  • Complete process for predicting circRNAs expression levels

a). Model training and feature selection( Used for obtaining rank of importance and select the best features)

CMD:

	circscan --exp-fs -t <cell_type> -m <model> -s/-r <seed>/<run (1-5)> -n <cores> [ --RM (remove outlier) ]
	# "-n": used for models training by parellel
	# "-s": used for random sample training set ("seed" mode, multiple traning and reproducibility )
	# "-r": used for 5 independent CV run ("run" mode)
	# "--RM": whether remove outlier data points

	e.g.:
	circscan --exp-fs -t K562 -m rf -s 111 -n 8
	or:
	circscan --exp-fs -t K562 -m rf -r 1 -n 8

Generate models and R data file "K562_rf_FS_exp.RData", log file "K562_rf_FS_exp.out", result of model performance "GM12878_rf_perf_test_reg.txt" with model evaluation of feature selection

b). Model performance evaluation (testing data)

CMD:

	circscan --exp-eval -t <cell_type> -m <model> -n <cores>
	# "-n": used for models training by parellel

	e.g.:
	circscan --exp-eval -t K562 -m rf -n 8

Generate models and R data file "K562_rf_pred_exp_all.RData", log file "K562_rf_eval_cor.out", result of model performance "K562_rf_eval_test_exp_perf" with model evaluation in testing data, results of predicted expression levels "K562_rf_obs_pred_exp_test.all", "K562_rf_obs_pred_exp_train.all"

c). Predict expression

CMD:

	circscan --exp-pred -t <cell_type> -m <model> -n <cores> -sf < all/select_fea_list (comma separated)> [ --RM (remove outlier) ]
	# "-n": used for models training by parellel
	# "-sf": select fea list (according to results of feature selection)
	# "--RM": whether remove outlier data points

	e.g.:
	circscan --exp-pred -t K562 -m rf -n 8 -sf all

Generate predicted circRNAs expression file "K562_rf_train_pred_exp"



All data are avaliable in "$PKG_DIR/sample" for performing and replicating the works in this Document.

circscan's People

Contributors

johnlcd avatar

Stargazers

avatar avatar avatar avatar avatar avatar

Watchers

avatar avatar

Forkers

studentyaoshi janewangsmile

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google ❤️ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.