Mutation counts from secondary alignments?

[physnano]

Hi Danny,

I have a question regarding how rf-count operates with alignments/SAM flags. Based on output from rf-rctools view I am seeing fewer (T->C) mutations {rfcount_7sK.jpg} (no positions > 1 mutation) than I would expect based on what I see with IGV (>9 positions with >10 T->C mutations). {IGV_7sK.jpg} When I look at the SAM flags for these particular alignments there is a preponderance of secondary alignments (~704) compared to primary alignments (~56) {sam_file_flag_breakdown_7sk.jpg} suggesting that rf-count is not counting these mutations from secondary alignments. However the coverage output does appear to be correct. Based on the option in rf-count to only consider primary alignments ("-pn") I would not have expected this default behavior. Can you comment on rf-count handling of secondary alignments?

{IGV_7sK.jpg}:

{rfcount_7sK.jpg}:

{sam_file_flag_breakdown_7sk.jpg}:

Thank you again and best regards,

Will

[dincarnato]

Hi Will,

As you can see from the number of options available in rf-count, there are many reasons a mutation might be ignored/discarded.

Can you please give me an example BAM file, corresponding Fasta, and command line you executed?

Danny

[dincarnato]

Also, please make sure you git pull the latest version as there was a bug in the "--only-mut" analysis that is now fixed.

[dincarnato]

Hi Will,

any update here?

Cheers, Danny

[physnano]

Hi Danny,

7SK seems to be a particularly problematic example of what I have attempted due to multiple transcripts/pseudogenes for this ncRNA. In this case, primary alignments were contributing to the coverage at secondary locations (but not the mutation counts, even in the case without the primary only option, or maybe I am interpreting the output incorrectly). Also it would be nice to have the coverage for each transcript reflected in not only the .rc file but also in the "raw_counts" output. That being said I think I have a handle on things now :) .

Overall thank you so much for your help and for developing this amazing package.

Best,

Will

[dincarnato]

Hi Will,

if you can provide me with an example BAM file, and explain me in detail the issue, I can maybe adjust things in the framework. You can send it to me via email.

Cheers, Danny