0 superenhancers identified

Issue #43 new
Former user created an issue

I ran the example files, and everything worked well, and 307 SEs were identified.

I am running ROSE now on H3K27ac ChIP-Seq, and after fixing GFF format issues, the script completed, but did not identify any SEs (gives graph as straight line at 0). It appears the Cutoff used was 0. All the expected output files exist. In the "AllEnhancers.table.txt" output file (see below), the NUM_LOCI column only shows values of 1, the CONSTITUENT_SIZE is much smaller than the example (first lines around 300, and ordered in increasing value), and the FN.bam columns have values of 0, which appears to be different from the example. Things I suspect may be wrong, but cannot determine (or could be something different): 1) Data could be of low quality and SEs cannot be determined. I used fastqs from SRA, so the ChIP-Seq was not performed by us. But the number of reads and quality metrics appears OK (fastqc and MACS), and the bams appear to have strong peaks with a few hundred overlapping reads in some loci (as viewed in IGV). What quality metrics should be met for ROSE to work? 2) Bams were not located in the ROSE folder, but were in the same folder as the matching index files. The bams were sorted, so that shouldn't be the issue. Maybe part of the script needs the bams to be located within the ROSE folder? (I'm currently re-running with files located within ROSE folder to test this). 3) When the script ran, the progressive counting of lines mapped to GFF precedes the R script, where in the example it occurs after. Is this order important? What would make it run out of sequence? 4) There is still some GFF formatting issue? I put the GFF file as an attachment. I used: awk 'BEGIN {OFS="\t"} {print "chr"$1, $4, "", $2, $3, "", ".", "", $4}' iFTSEC33rep1_peaks.bed >iFTSEC33rep1_peaks.gff to convert the MACS1.4.2 peak bed file to the GFF format required for ROSE. 5) I'm running locally on my Mac (10.13.6) in terminal.

bash command and output in attached file.

"AllEnhancer.table.txt" head: REGION_ID CHROM START STOP NUM_LOCI CONSTITUENT_SIZE SRR1983932.bam SRR1983936.bam enhancerRank isSuper 1_MACS_peak_33416_lociStitched chr21 40317690 40317987 1 298 0 0 1 0 1_MACS_peak_4406_lociStitched chr1 204616569 204616890 1 322 0 0 2 0 1_MACS_peak_19938_lociStitched chr16 28126999 28127339 1 341 0 0 3 0 1_MACS_peak_28228_lociStitched chr2 69326638 69326978 1 341 0 0 4 0 1_MACS_peak_47562_lociStitched chr6 143607479 143607830 1 352 0 0 5 0 1_MACS_peak_22401_lociStitched chr17 45197900 45198280 1 381 0 0 6 0 1_MACS_peak_53595_lociStitched chr9 8391336 8391719 1 384 0 0 7 0

Comments (1)

  1. Deniz Törüner

    I have the same problem too. Rose identifies 0 super-enhancers. I would be very grateful if you could let me know if it eventually worked and what made it work. Thanks in advance.

  2. Log in to comment