Visualization tool to compare distribution profiles of NGS reads?

[Kennels]

Hello,

Does any one know of a (free) software that is able to normalize aligned NGS data of different samples (e.g. in .bam format) to enable comparison of their read distribution profile across a reference sequence? That is, I want to compare not only the distribution profile of reads mapped onto the reference but also compare their relative abundance to one another, simultaneously.

I've tried Integrated genome viewer, but would like to try other software as well.

Thanks for any information!

Ken

[rururara]

i try artemis and tablet but i failed to view in IGV even though I already sorted and index the BAM file.
any tips can you share with me to view in IGV because so many times i tried but still can't view from IGV.

[maubp]

Tablet only loads one BAM file at a time, so to do comparisons with Tablet you'd need to have one Tablet window open per BAM file, and manually line up the views. Not ideal.

I would use Artemis with the BamView plugin. This can load a reference sequence with multiple BAM files, which you can color code. You can show a coverage plot as well as the raw reads. The user interface is a bit cumbersome, but it works.

http://www.sanger.ac.uk/resources/software/artemis/

[simonandrews]

If your reference sequence is an assembled genome then you can do this in SeqMonk (which comes from our group)

[honey]

Seq monk

Hi Simon,

Can we plot RPKM/FPKM values for visualization in Seqmonk?

[simonandrews]

Can we plot RPKM/FPKM values for visualization in Seqmonk?

Kind of. In the current release you can calculate values which are adjusted to the size of the largest dataset rather than per million reads. These are equivalent to RPKM but scaled up (or down) by a constant factor. In the next release there will be an option to correct per million reads to get RPKM directly.

The reason we don't do RPKM at the moment is that if you log transform your data (which makes a lot of sense when viewing most types of expression data), then you often end up with a lot of negative values which make the plots more difficult to interpret than if everything is on a positive scale.

[Kennels]

Hi rururara

I've been viewing the .tdf file with the 'count' function from igvtools.
did you try to generate the .tdf coverage file with igvtools, or just tried to view along with the bam file? Sometimes if the reference is too large it won't display it until you zoom in quite abit.
Let me know what you did specifically and i'll get back to you with more details.

thanks everyone else

[Jim Robinson]

If you create a coverage (tdf) file with igvtools you can normalize it when viewing by changing a preference. Specifically, select View > Preferences, click the "Tracks" tab, and select the last checkbox "Normalize Coverage Data". This will scale each value by 1,000,000 / total mapped count.

I realize this is not a very intutitve way to turn this feature on, its on my list to fix.

Also on my list is implementing RPKM and FPKM, both in igvtools and "on-the-fly" in igv. I understand the concept but was here on seqanswers hoping to to find a definition that acounts for alternative isoforms. Is the normal thing to take the union of all exons for all transcripts, or what?

Jim

[Kennels]

Jim,

Not sure but these might be relevant to your question:

Accurate quantification of transcriptome from RNA-Seq data by effective length normalization
http://nar.oxfordjournals.org/content/39/2/e9.full

Prediction of alternative isoforms from exon expression levels in RNA-Seq experiment (they used IGV in this one)
http://nar.oxfordjournals.org/content/38/10/e112.full

[honey]

Igv

Hi Jim,
I was wodnering have you implemented changes in IGV so as to plot FPKM/RPKM normailzed values?
Thanks.