Fatal error, distance matrix is too sparse

[pweigele]

How can I rectify output from command:
muscle -super7 ThyAs.mega -distmxin ThyAs.distmx -reseek -output ThyAs.afa
producing
---Fatal error---
Distance matrix too sparse
?

[rcedgar]

This means that the distance matrix does not have enough information to build a guide tree. Try using the -sensitive or -verysensitve option for reseek when making the matrix. If that doesn't work either, the structures are highly divergent and you probably won't get a meaningful MStA.

[pweigele]

Thank you, that makes sense! Perhaps off topic then, but I am working with a directory of ~800 models-- do you know of a way to cluster them from sequences in .pdb file (rather than going back to the .fasta and manually removing files from the directory)?

…

On Thu, Nov 28, 2024, 10:20 AM Robert Edgar ***@***.***> wrote: This means that the distance matrix does not have enough information to build a guide tree. Try using the -sensitive or -verysensitve option for reseek when making the matrix. If that doesn't work either, the structures are highly divergent and you probably won't get a meaningful MStA. — Reply to this email directly, view it on GitHub <#16 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AICFWVXHCGJOKH6B72HWCGD2C4YDXAVCNFSM6AAAAABSVKXLASVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDKMBWGM2TENJWGU> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[rcedgar]

Something like this

usearch -cluster_fast pdb.fasta -id 0.5 -centroids centroids.fasta
grep '^>' centroids.fasta | tr -d '>' > pdbids.txt
for label in `cat pdbids.txt`
do
  mv -v $label.pdb ../some_other_directory
done

[pweigele]

Thank you for the suggestion! Btw, the version of muscle5 I am running was installed with conda. I did as you suggested above, but still had the following error: (muscle) ***@***.*** ~]$ muscle -super7 ThyAs.mega -distmxin ThyAs.distmx -reseek -output ThyAs.afa muscle 5.3.linux64 [-] 1584Gb RAM, 56 cores Built Nov 11 2024 08:05:12 (C) Copyright 2004-2021 Robert C. Edgar. https://drive5.com [super7 ThyAs.mega] WARNING: Duplicate sequences found Input: 622 seqs, avg length 288, max 642, min 99 00:00 26Mb Reading dist mx (reseek format)...0 pair-wise distances

…

---Fatal error--- Distance matrix too sparse (muscle) ***@***.*** ~]$

On Fri, Nov 29, 2024 at 1:10 PM Robert Edgar ***@***.***> wrote: Something like this usearch -cluster_fast pdb.fasta -id 0.5 -centroids centroids.fasta grep '^>' centroids.fasta | tr -d '>' > pdbids.txt for label in `cat pdbids.txt` do mv -v $label.pdb ../some_other_directory done — Reply to this email directly, view it on GitHub <#16 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AICFWVQVIXL4WJTT6SYGNPT2DCUZFAVCNFSM6AAAAABSVKXLASVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDKMBYGIZDMNZZGE> . You are receiving this because you authored the thread.Message ID: ***@***.***>