rpkm/tpm rather than median/mean reads

[mhassan]

am plotting two samples on using the same settings on the the example_run.sh, except the gtf file (which is mine). However, while the y axis scale on the first automatically readjusts to median read count, panel 2 is stuck to one. Is there a way to make sure the scales on both panels readjust automatically.
Also is it possible to visual rfkm/tpm rather than median/mean reads?

[dgarrimar]

Hi @mhassan, regarding your first question, could you provide a sample plot to better diagnose what's going one? In principle both scales should be adjusted as expected. Regarding your second question, it is currently not possible. But how would you do that? Think that we are showing the signal for a particular region, while RPKM/TPM would correspond to the whole gene. Therefore it is the signal track what is more informative, right? @abreschi @emi80

[mhassan]

Hi, I’ve figured out the problem with the scales. Regarding the TPM, I guess my question was informed by the fact that I was using this tool to plot coverage on whole transcript, not just a splicing event, in the same way that sashimi-plot does.

On Mon, Oct 29, 2018 at 9:00 AM Diego Garrido-Martín < ***@***.***> wrote: Hi @mhassan <https://github.com/mhassan>, regarding your first question, could you provide a sample plot to better diagnose what's going one? In principle both scales should be adjusted as expected. Regarding your second question, it is currently not possible. But how would you do that? Think that we are showing the signal for a particular region, while RPKM/TPM would correspond to the whole gene. Therefore it is the signal track what is more informative, right? @abreschi <https://github.com/abreschi> @emi80 <https://github.com/emi80> — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#10 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAVQQlkZ7dsjdsCtai4JkLVvmHoDKHBIks5upsOmgaJpZM4XwB41> .

-- Sent from Gmail Mobile

[dgarrimar]

Do you think something like this (pseudo RPKM) would be useful?

where query length is the length of the aligned query sequence.

[emi80]

Hi @dgarrimar, @mhassan
I think this could be done. The only problem I see is that getting the total number of reads from the BAM would slow down the process quite a bit with big files and with many files it might end up being very slow.

We might look into multiprocessing or other ways to improve this if we really need it.

[stale]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

[ckovalak]

Hi @emi80
Just wanted to bump this because I think having normalized read counts would be extremely helpful when comparing across datasets of varying depth. To keep it simple, maybe just an option to report RPM (reads per million) rather than a raw read count? Total library size could potentially be included as an additional column in the input_bams.tsv as this value would not change and would save considerable time calculating this number each plot.

[bdgsilva]

Hi @emi80 @dgarrimar,
Also wanted to further support the implementation of normalized read densities as it would be extremely useful when comparing samples with different sequencing depths. @ckovalak suggestion is a good option and I believe it was how MISO implemented this same feature.