1
$\begingroup$

I've been trying to run mmseqs2 on a few metagenomic assemblies and despite my best efforts in reading the wiki and playing with parameters, the process is taking over a day.

In their paper they claim to be 2 - 18x faster than CAT. I ran CAT on the same datasets and they took at most a few hours. So I can only assume that the long run time for mmseqs is down to user error (i.e. me).

I am currently running the commands (general format):

mmseqs createdb metagenome1_scaffolds.fasta metagenome1_query_DB
mmseqs taxonomy metagenome1_query_DB /home/mmseqs2/NR/NR metagenome1_mmseqs_taxonomy.txt metagenome1_temp --tax-lineage 1 --orf-filter 1 --threads 24

Could someone please let me know where I'm going wrong or what parameters I can tweak to get this running in the expected time frame?

Improve this question
$\endgroup$
2
  • $\begingroup$ Have you contacted the developers about this? Do any of these issues match the problem you're having? $\endgroup$
    – gringer ♦
    Commented Dec 2, 2023 at 8:41
  • 1
    $\begingroup$ Hi there! I actually got in running in ~4 hours but then the results are split into 24 semi-corrupted files (I used 24 threads). I have since contacted the developers and will post a solution once I get a response and figure it out :) $\endgroup$
    – Rainman
    Commented Dec 3, 2023 at 22:28

1 Answer 1

Reset to default
1
$\begingroup$

After much googling about I realised that I was missing a step! Here are the commands I used to get it running:

mmseqs createdb metagenome1_scaffolds.fasta metagenome1_query_DB
mmseqs taxonomy metagenome1_query_DB /home/mmseqs2/NR/NR metagenome1_mmseqs_taxonomy.txt metagenome1_temp --tax-lineage 1 --orf-filter 1 --threads 24 --split-memory-limit 500G 
mmseqs createtsv metagenome1_DB metagenome1_mmseqs_taxonomy.txt metagenome1_mmseqs_taxonomy_final.txt

The split-memory-limit flag sped up the process and the final line put all the split files back together

Improve this answer
$\endgroup$
1
  • 1
    $\begingroup$ Thanks @Rainman $\endgroup$
    – M__ ♦
    Commented Dec 4, 2023 at 18:01

Not the answer you're looking for? Browse other questions tagged or ask your own question.