Elementolab/ChIPseeqerModel tools

From Icbwiki

Jump to: navigation, search

ChIPseeqer Model tools

These tools are included in ChIPseeqerModel, but can also be used separately.

Requires: ChIPseeqer (with PCA/GSL), Heatmap perl module, R.

1. Apply CS to get peak files

ChIPseeqer.bin -chipdir H3K9Ac/CHIP/ -inputdir H3K9Ac/INPUT/ -outfile h3k9Ac_peaks.txt
ChIPseeqer.bin -chipdir MTA3/CHIP/ -inputdir MTA3/INPUT/ -outfile MTA3_peaks.txt
...

2. Get promoter binding scores

perl $CHIPSEEQERDIR/SCRIPTS/CalcExtendedPeakScores.pl --peakfile=h3k9Ac_peaks.txt --d0=5000 --label=H3K9Ac > h3k9Ac_peaks.txt.PB
perl $CHIPSEEQERDIR/SCRIPTS/CalcExtendedPeakScores.pl --peakfile=MTA3_peaks.txt --d0=5000 --label=MTA3 > MTA3_peaks.txt.PB
...

3. Combine multiple promoter scores

expression_concatenate_matrices.pl */*.PB > MatPromBind.txt

4. PCA

ChIP_PCA -chipdata MatPromBind.txt -center 0 -scale 1

(NOTES: gsl library is needed

ChIP_PCA is in BCELLS)

This will create 3 files, ending in .evec, .eval and .proj

5. visualize eigenvectors and eigenvalues

draw_expression_heatmap.pl --matrix=MatPromBind.txt.evec --minmax=0.25 --clustcols=1 \
  --distance=euclidean --cmap=/home/ole2001/PROGRAMS/HEATMAP/colormaps/cmap2.txt
plotEval.pl MatPromBind.txt.eval 

6. Generate RNA-seq RPKM

Simple format: Transcript_ID RPKM_value

ID	LY1	LY7	CB	NB
NM_021648	13.6	14.4	14.9	10.5
NM_020351	0.0	0.0	0.0	0.0
NM_001112734	1.4	1.6	3.6	3.4
NM_002697	21.9	28.1	12.7	7.0
NM_002656	0.1	0.0	0.7	0.8

7. Linear regression

perl ~/PROGRAMS/BCELLS/CorrelatePCADataAndExpression.pl --evec=PromBindMatAllNOH_DNAMETHnomissing.txt.proj --expfile=/panda_scratch_miro/ole2001/MELNICKRNASEQ/LY1_LY7_CB_NB_RPKM.txt --resp="log(LY1+1)"
Personal tools