search

NKI-GCF / XenofilteR

Filtering of PDX samples for mouse derived reads
GNU General Public License v3.0
27 stars 7 forks source link

Recommended parameters for 110 bp paired-reads #25

Closed ZReplicant closed 1 year ago

ZReplicant commented 1 year ago

Hello!

Thanks for developing this tool. I tried running XenofilteR, and it worked without any issues. However, I used the basic command mentioned in the readme: XenofilteR(sample.list, destination.folder = in_path, bp.param = bp.param, output.names) Even though my reads are 110 bp (paired-end, RNA-seq DEG-analysis experiment).

I thought I could change the parameters like it was written in the readme. Would you please recommend the values for MM_threshold and Unmapped_penalty as I was not able to google any clues?

Thank you.

RoelKluin commented 1 year ago

Dear Zhanna,

For pe 110bp I'd recommend paired-end adapter trimming beforehand, and maybe quality trimming, if e.g. fastqc shows quality declines with length. Tools: Seqpurge, or cutadapt if quality trimming is required. Soft-clipped bases are counted as mismatches. Thresholds should still be ok. You could elevate the mismatch threshold and unmapped penalty to 8 and 12, or disable the first by setting it to readlength.

Best regards, Roel

On Wed, 15 Mar 2023, 19:07 Zhanna, @.***> wrote:

Hello!

Thanks for developing this tool. I tried running XenofilteR, and it worked without any issues. However, I used the basic command mentioned in the readme: XenofilteR(sample.list, destination.folder = in_path, bp.param = bp.param, output.names) Even though my reads are 110 bp (paired-end, RNA-seq DEG-analysis expreiment).

I thought I could change the parameters like it was written in the readme. Would you please recommend the values for MM_threshold and Unmapped_penalty as I was not able to google any clues?

Thank you.

— Reply to this email directly, view it on GitHub https://github.com/NKI-GCF/XenofilteR/issues/25, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAERVP5NS6PJ4LBZWPXZYGTW4IAO5ANCNFSM6AAAAAAV4GGOKA . You are receiving this because you are subscribed to this thread.Message ID: @.***>

ZReplicant commented 1 year ago

Dear Roel,

Thank you very much for the fast answer. I did perform adapter and quality trimming beforehand. Then, I will probably stick with the default parameters. Will raising the mismatch threshold and unmapped penalty to 8 and 12 improve specificity in my case?

Best regards, Zhanna

RoelKluin commented 1 year ago

I'd say raising thresholds can increase sensitivity, but specificity can only decrease.

ZReplicant commented 1 year ago

Ok, thank you again. Your answers really helped!