For bam files aligned to transcriptome, I am getting the error 'Unknown sequence name'?

[rnaomics]

Hi,

I have bam files aligned to the transcriptome. I used the per-base base depth function on those bam files. I am getting the following error.

thread '' panicked at 'Fetched reference sequence: Unknown sequence name: NM_001011874.', src/commands/base_depth.rs:275:30
note: run with RUST_BACKTRACE=1 environment variable to display a backtrace

[sstadick]

I believe this is because the bam header isn't up to date with the refs used? Could you share a small bam file / the refs your'e using?

Sorry for the long delay in reply!

[rnaomics]

Hi The tool fails when it come across this reference sequence Chr_ENSMUST00000031779.10 and I got this error. [image: unnamed.png] Best regards, Mohamed

…

On Tue, Jul 22, 2025 at 5:02 PM Seth ***@***.***> wrote: *sstadick* left a comment (sstadick/perbase#75) <#75 (comment)> I believe this is because the bam header isn't up to date with the refs used? Could you share a small bam file / the refs your'e using? Sorry for the long delay in reply! — Reply to this email directly, view it on GitHub <#75 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AQ6I2OH7U63Q6W4FCWB36G33JY76FAVCNFSM6AAAAACCC73WCOVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZTCMBSHA4DANJRGY> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[rnaomics]

Hi Upon inspection, I noticed a clear pattern: the crash occurs whenever `perbase` encounters certain transcript IDs in the BAM, for example: Chr_ENSMUST00000031779.10 Chr_ENSMUST00000034203.10 Chr_ENSMUST00000045884.10 - These transcripts **exist in the FASTA**, so the issue is not a missing sequence. - The pattern suggests that the problem is systematic for transcripts whose names contain trailing bytes or non-printable characters in the Rust strings extracted from the BAM header. - Removing individual transcripts from the BAM avoids the crash temporarily, but the next transcript triggers the same panic, so this is not a practical workaround. From my investigation, it appears that `pos.ref_seq` in `base_depth.rs` may contain hidden/trailing characters, which causes `bio::io::fasta::IndexedReader::seq()` to fail even when the sequence exists. Thank you for your time and for developing `perbase` — it’s a very useful tool!

…

On Tue, Jul 22, 2025 at 5:02 PM Seth ***@***.***> wrote: *sstadick* left a comment (sstadick/perbase#75) <#75 (comment)> I believe this is because the bam header isn't up to date with the refs used? Could you share a small bam file / the refs your'e using? Sorry for the long delay in reply! — Reply to this email directly, view it on GitHub <#75 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AQ6I2OH7U63Q6W4FCWB36G33JY76FAVCNFSM6AAAAACCC73WCOVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZTCMBSHA4DANJRGY> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[sstadick]

@rnaomics Thanks for the added details! That helps a lot!

I'm away for the holidays, but will look in on this when I get back!