MMseqs2 icon indicating copy to clipboard operation
MMseqs2 copied to clipboard

mmseqs taxonomy Error: orf filter prefilter died / Error: Search died

Open jameslz opened this issue 2 years ago • 1 comments

Expected Behavior

k-mer similarity threshold: 145 Starting prefiltering scores calculation (step 1 of 1) Query db start 1 to 469186 Target db start 1 to 579233 [=================================================================] 100.00% 469.19K 3s 382ms

6.094516 k-mers per position 435 DB matches per sequence 0 overflows 0 queries produce too many hits (truncated result) 0 sequences passed prefiltering per query sequence 0 median result list length 464221 sequences with 0 size result lists Time for merging to orfs_pref: 0h 0m 0s 75ms munmap_chunk(): invalid pointer tmp/12438089539006473617/taxonomy_tmp/18301032544654743241/taxpercontig.sh: line 66: 3001403 Aborted (core dumped) $RUNNER "$MMSEQS" prefilter "${ORFS_DB}" "${TARGETDB_IDX}" "${TMP_PATH}/orfs_pref" ${ORF_FILTER_PREFILTER} Error: orf filter prefilter died Error: Search died

Steps to Reproduce (for bugs)

mmseqs easy-taxonomy -s 7.5 C10.fasta ../db/refseq-prot/mmseqs/refseq_viral result tmp --lca-mode 2

Context

I try to annotate virus contig with refseq database uing mmseqs easy-taxonomy, I get errors below, When I use the same command with tmp files, I can get the taxonomy annotation files.

But I use '--orf-filter 0', I get no errors.

Your Environment

mmseqs version 45111b641859ed0ddd875b94d6fd1aef1a675b7e mmseqs2-sse4-r13

jameslz avatar Apr 19 '22 03:04 jameslz

I encountered the orf filter prefilter died issue today. I suspect it might be due to insufficient memory available to the program, even though there was no indication of memory shortage. According to the MMseqs2 documentation, the program should automatically analyze the available memory and split the data accordingly to meet memory requirements. However, I suspect the program did not correctly identify the maximum memory I allocated for my task submitted to HPC (via qsub), and it may have mistakenly considered the total available memory in HPC as the memory accessible for the current task. I resolved this issue by adding the parameter --split-memory-limit 80G. Hope this helps you.

SilentGene avatar Jan 09 '24 07:01 SilentGene