.bam to .wig conversion

[kalidaemon]

I'm trying do some sanity checks by putting my TopHat output on the UCSC browser. In order to do this, I need to convert my accepted_hits.bam files to .wig files. Has anyone tried to do this? I know you can do this with SAMtools for .sam files and I was wondering if you can do the same for .bam. If not, what other tools do people use? Also, older TopHat output generated .wig files. Is there any way to get newer builds to do this? Thanks!

[natstreet]

I've resorted to converting the .bam file to -sam and using the wiggles program that comes with tophat to create the .wig files but this is hugely inefficient. I really wish they had kept the creation of wig files as standard or that wiggles could be updated to also accept .bam files as input. Then again, I also wish they hadn't dropped support for GFF3 in favour of GFF2!

[dawe]

I'm trying do some sanity checks by putting my TopHat output on the UCSC browser. In order to do this, I need to convert my accepted_hits.bam files to .wig files. Has anyone tried to do this? I know you can do this with SAMtools for .sam files and I was wondering if you can do the same for .bam. If not, what other tools do people use? Also, older TopHat output generated .wig files. Is there any way to get newer builds to do this? Thanks!

You can try genomeCoverageBed from bedtools suite, it's able to read bam and generate bedgraph files.

d

[RockChalkJayhawk]

I've resorted to converting the .bam file to -sam and using the wiggles program that comes with tophat to create the .wig files but this is hugely inefficient. I really wish they had kept the creation of wig files as standard or that wiggles could be updated to also accept .bam files as input. Then again, I also wish they hadn't dropped support for GFF3 in favour of GFF2!

I think the reason for dropping wiggle tracks is because they are massive and the server gets bogged down with trying to import all that data. Besides, BAM files can be visualized directly on UCSC - if you can put them on a web-accessible server.

[mgogol]

If anyone's using bedtools genomeCoverageBed for this, don't forget to use the -split parameter if you're using it with tophat bam files, otherwise the junction reads get stretched across the introns weirdly.



The commands I ended up using were

genomeCoverageBed -split -bg -ibam accepted_hits.sorted.bam -g dm3.chrom.sizes > accepted_hits.bedgraph
wigToBigWig accepted_hits.bedgraph dm3.chrom.sizes myfile.bw

[anagari]

Prob with visualization even after using split

Hi,

I used the command as you have suggested to split the reads over the splice junctions:
./genomeCoverageBed -split -bg -ibam ip_sorted.bam -g genome_hg19.txt > accepted_hits.bedgraph

#genome_hg19.txt has the chromosome sizes for hg19.

But, when I upload the bedgraph onto Genomebrowser, it still shows stretch of reads between peaks.

Could you suggest where the problem might be?

Thanks in advance!

[shenr]

This issue bothered me for a while too when I used bedtools v2.13, in which -split didn't work.

The fix is actually simple: just download the new version bedtools (v2.16.2) and re-run it.

Hi,

I used the command as you have suggested to split the reads over the splice junctions:
./genomeCoverageBed -split -bg -ibam ip_sorted.bam -g genome_hg19.txt > accepted_hits.bedgraph

#genome_hg19.txt has the chromosome sizes for hg19.

But, when I upload the bedgraph onto Genomebrowser, it still shows stretch of reads between peaks.

Could you suggest where the problem might be?

Thanks in advance!

[anagari]

@Shenr: Thanks a lot!! Updating bedtools to 2.16 did work!!