What is the corresponding caculation formula for TPM? And could SUPPA be used in long-read RNA-seq?

[xyz202310]

Hi, thanks for your great work!
Could you please tell me the corresponding caculation formula for TPM? I noticed that common TPM caculation formula need to use gene length to correct gene expression. I'm not sure if TPM in SUPPA is corrected based on gene length.
And could SUPPA be used in long-read RNA-seq? If I use SUPPA to analyze long read RNA-seq data, do I need to replace TPM with count?
Thank you!

[EduEyras]

Hi, thanks for your message

On Fri, 23 May 2025 at 16:17, xyz202310 ***@***.***> wrote: *xyz202310* created an issue (comprna/SUPPA#214) <#214> Hi, thanks for your great work! Could you please tell me the corresponding caculation formula for TPM? I noticed that common TPM caculation formula need to use gene length to correct gene expression. I'm not sure if TPM in SUPPA is corrected based on gene length.

SUPPA does not calculate the TPM. You need to pass to SUPPA the TPM calculated from a different program. You're probably referring to methods like tximport that takes TPM values to calculate the length-normalised gene-based read counts to perform differential gene expression analysis from TPM values using standard tools like DESeq2, limma, etc. The gene-correction that takes place in SUPPA has to do with the multiple-test correction for differential splicing, and it does not involve transcript or gene lengths, it's all about the fact that events within a gene are likely to be compared to each otther.

And could SUPPA be used in long-read RNA-seq? If I use SUPPA to analyze long read RNA-seq data, do I need to replace TPM with count?

Yes, SUPPA can use any TPM values, calculated from any data source, including long-read data. In long-read data one read is one transcript count, so you may not need to take length into account, and TPM can be calculated simply as read counts per million. I hope this helps Eduardo

…

Thank you! — Reply to this email directly, view it on GitHub <#214>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ADCZKBZ6QZZQ7SOYDHJUV2L2724O5AVCNFSM6AAAAAB5X3MHJ6VHI2DSMVQWIX3LMV43ASLTON2WKOZTGA4DKMZRGA2DSOI> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>