Regarding the PR calculation script issue

[QY6616]

@martin-steinegger
When I use the following command: ./bench.fdr.noselfhit.awk scop_lookup.fix.tsv <(cat mmseqs2.txt) > test.rocx to calculate the PR value for mmseqs2, the output is as shown in the figure; however, when calculating the PR value for dali, everything works fine. Could you please explain why this is happening?

[martin-steinegger]

How does your MMseqs2.txt file look like?

[QY6616]

您的 MMseqs2.txt 文件是什么样的？

@martin-steinegger I downloaded it from https://wwwuser.gwdguser.de/~compbiol/foldseek/scop.benchmark.result.tar.gz without making any modifications.

[martin-steinegger]

I think the input should be sorted by the score (last column). Does <(sort -k12,12nr mmseqs2.txt) help?

[QY6616]

我认为输入应该按分数（最后一列）排序。有<(sort -k12,12nr mmseqs2.txt)帮助吗？

@martin-steinegger Awesome! The issue is resolved. So, which column should foldseekTM.txt and foldseek.txt be sorted by?

[martin-steinegger]

Foldseek should be sorted by the 12th column (score) and Foldseek-TM by the 11th, which should be the the average TMscore ((qTM+tTM)/2)

[QY6616]

@martin-steinegger Thank you very much for your help. I have successfully reproduced the relevant curves. I have a few more questions: Where can I download the "scop-fasta" file mentioned in the runMMseqs.sh script? Do the files "scop.file.full.name.list" and "cle.file.list" refer to the "scop_lookup.fix.tsv" file? I look forward to your answer. Thank you very much.

[martin-steinegger]

The scop-fasta folder contains protein sequences for each SCOP domain, each stored as a single .fasta file (attached below). The files cle.file.list and scop.file.full.name.list include the domain names, which correspond to the names listed in scop_lookup.fix.tsv.

scop-fasta.tar.gz

[QY6616]

@martin-steinegger That is to say, can I use scop_lookup.fix.tsv to replace cle.file.list and scop.file.full.name.list?

…

---- Replied Message ---- | From | Martin ***@***.***> | | Date | 04/11/2025 17:33 | | To | ***@***.***> | | Cc | ***@***.***>***@***.***> | | Subject | Re: [steineggerlab/foldseek-analysis] Regarding the PR calculation script issue (Issue #11) | The scop-fasta folder contains protein sequences for each SCOP domain, each stored as a single .fasta file (attached below). The files cle.file.list and scop.file.full.name.list include the domain names, which correspond to the names listed in scop_lookup.fix.tsv. scop-fasta.tar.gz — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: ***@***.***> martin-steinegger left a comment (steineggerlab/foldseek-analysis#11) The scop-fasta folder contains protein sequences for each SCOP domain, each stored as a single .fasta file (attached below). The files cle.file.list and scop.file.full.name.list include the domain names, which correspond to the names listed in scop_lookup.fix.tsv. scop-fasta.tar.gz — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: ***@***.***>