Read length determination

[Nicolai-vKuegelgen]

While looking through the output & code of riboseqc, I've noticed that the determination of read_lengths is based only on the mapped part of each read, while soft-clipped parts of the read are ignored.

This means that sequence/read with 31nt but only 29nt mapped & 2nt softclipped will (fasley?) be regarded as a 29nt read.

Additionally, this probably also affects the P-site assignment of such reads, if the softclipped part is at the start of the read.

I'm not sure about the reasoning behind this approach, since the reads produced by Riboseq should all come from ribosome protected fragments and the full length should be counted indepent of whether it was fully/efficiently mapped.

[Xue-yd]

Hi, I've also encountered this problem recently when using RiboseQC. How did you finally solve it?
Thanks a lot if you could give me some advice.

While looking through the output & code of riboseqc, I've noticed that the determination of read_lengths is based only on the mapped part of each read, while soft-clipped parts of the read are ignored.

This means that sequence/read with 31nt but only 29nt mapped & 2nt softclipped will (fasley?) be regarded as a 29nt read.

Additionally, this probably also affects the P-site assignment of such reads, if the softclipped part is at the start of the read.

I'm not sure about the reasoning behind this approach, since the reads produced by Riboseq should all come from ribosome protected fragments and the full length should be counted indepent of whether it was fully/efficiently mapped.

[Nicolai-vKuegelgen]

Hey,
I think I've never found a solution for that issue within the scope RiboseQC. If you are looking for other solutions I could try to check my old scripts from back then.

[Xue-yd]

Thanks for your timely reply！
I think I know what the problem is. The data I work with is from SRA database. I found the process pipeline of the lab where the data comes from. The softclipped sequence might originate from UMIs that are not removed efficiently. The generation of this UMIs may be due to different protocols. After I processed the data according to their pipeline, the problem was solved.
Best regards!