all-vs-all BLASTP returns Error: generate_output: target with no hsps

[biotovarx]

Hello,

I am trying to perform an all-vs-all BLASTP between a set (4716 in total) of predicted proteins. I ran the following command to generate the database:

diamond makedb \
  --in predicted_proteins.faa \
  --db protein_db

I then run the following command with the predicted proteins fasta file as the query.

diamond blastp \
    --query predicted_proteins.faa \
    --db protein_db \
    --out blastp.tsv \
    --outfmt 6 \
    --evalue 1e-5 \
    --max-target-seqs 10000 \
    --query-cover 50 \
    --subject-cover 50

The above diamond command ends with the following error though:

Computing alignments... Error: generate_output: target with no hsps.

With an all-vs-all BLASTP, shouldn't each protein query have at least 1 hsp with itself in the database?

[bbuchfink]

Hi, would it be possible to make your input file available to me so I can reproduce the problem?

[DuncanSussfeld]

Hello, I am encountering the exact same error trying to perform a regular (not all-vs-all) alignment with the following command:

diamond blastp \
--query composite_domains.faa \
--db cyano_derep.dmnd \
--out /home/duncan/domains_vs_cyano.diamond.out \
--threads 30 --evalue 0.00001 --id 30 --query-cover 80 --more-sensitive -k 100 \
--outfmt "6 qseqid sseqid qlen slen pident evalue bitscore qstart qend sstart send length qcovs"

I am happy to share my input files with you if needed (note that the query takes up 163M of memory, and 1.4G for the database). Thanks

[bbuchfink]

@DuncanSussfeld If you could share those files with me, that would be great.

[zhangwenda0518]

In the process of all-vs-all BLASTP when I used orthofinder and orthomcl, I want to improve the speed through diamond, but I encountered the same problem。 **

diamond blastp --db goodProteins --query goodProteins.fasta --out diamond.xml --outfmt 6 --sensitive --max-target-seqs 500 --evalue 1e-5 --id 20 --tmpdir /dev/shm Error: generate_output: target with no hsps. ** Did you have solved the problem? I used blast for all vs all blastp again,there was no errors, but the speed was very slow！

Thanks for your reply ！Best to you！

[bbuchfink]

This patch should fix the issue: a4338229300d693f59f42657596f5bf261c2630e It will be included in the next release.

[bbuchfink]

This should be fixed in v2.0.14.

[ZhangDengwei]

Has it been fixed? I am still suffering from this issue.

[bbuchfink]

What diamond version are you using?

[ZhangDengwei]

What diamond version are you using?

I tested v2.0.9, v2.0.14, and v2.0.15. None of them worked for me.

My file "cluster_288" is

bgc100078__5001:5210:- MKNQLNFEILSDEELQGINGGIRGTGKGLAAAMVSGAAMGGAIGAFGGPVGAIMGAWGGAVGGAMKYSI bgc45988__29:238:+ MKNQLNFNIVSDEELQGINGGIRGTGKGLAAAMVSGAAMGGAIGAFGGPVGAIMGAWGGAVGGAMKYSI

The version here I tested is 2.0.15, my commands are as follows

/data3/zhangdw/01.Toolkit/diamond/diamond makedb --in cluster_288 -d cluster_288

/data3/zhangdw/01.Toolkit/diamond/diamond blastp --threads 6 --db cluster_288.dmnd --outfmt 6 --query cluster_288 --id 50 --query-cover 90

No hits were found.

[bbuchfink]

Please clarify, you did not find hits or you got the error message as described in the title?

[ZhangDengwei]

Please clarify, you did not find hits or you got the error message as described in the title?

Did not find any hit, but no error or warning like the title was received. I am wondering why no hits were found as I performed the all-vs-all blastp.

[bbuchfink]

Try using --masking 0.

[ZhangDengwei]

--masking 0

It worked now. Many thanks!