BLAST+ exits with error exit status (2) when using nextflow - blast

Question

I'm using nextflow to analyse minION data. Blast+ terminates with error exit status (2), Command exit status:2 and Command output: (empty)
-HP-Z6-G4-Workstation:~/nextflow_pipelines/nf_pipeline/20221025_insect$ nextflow cat_working_nextflow.nf
N E X T F L O W ~ version 22.04.5
Launching `cat_working_nextflow.nf` [admiring_hopper] DSL1 - revision: 2916bc12af
executor > local (78)
[38/2d0584] process > concatinate (AIG363_pass_barcode01_0eb3c2c3_2.fastq) [100%] 38 of 38 ✔
[dd/3cabdf] process > fastqconvert (output.fastq) [100%] 38 of 38 ✔
[47/dab2cd] process > blast_raw (insect.fasta) [ 0%] 0 of 38
executor > local (78)
[38/2d0584] process > concatinate (AIG363_pass_barcode01_0eb3c2c3_2.fastq) [100%] 38 of 38 ✔
[dd/3cabdf] process > fastqconvert (output.fastq) [100%] 38 of 38 ✔
[47/dab2cd] process > blast_raw (insect.fasta) [ 2%] 1 of 37, failed: 1
Error executing process > 'blast_raw (insect.fasta)'
Caused by:
Process `blast_raw (insect.fasta)` terminated with an error exit status (2)
Command executed:
blastn -query insect.fasta -db /home/blast/nt_db_20221011/nt -outfmt 11 -out blastrawreads.asn -evalue 0.1 -numgnments 1
blast_formatter blastr-archive blastrawreads.asn awrea-outfmt 5 -out blastrawreads.xml
blast_formatter -archive blastrawreads.asn -outfmt "6 qaccver saccver pident length evalue bitscore stitle" -out blastrawreads_rt.tsv
sort -n -r -k 6 blastrawreads_unsort.tsv > blastrawreads.tsv
Command exit status:
2
Command output:
(empty)
Command error:
Warning: [blastn] Examining 5 or more matches is recommended
BLAST Database error: No alias or index file found for nucleotide database [/home/blast/nt_db_20221011/nt] in search path [/home/shaextflow_pipelines/nf_pipeline/20221025_insect/work/96/e885b7e53e1bcf30e33526265e9a3c::]
Work dir:
/home/nextflow_pipelines/nf_pipeline/20221025_insect/work/96/e885b7e53e1bcf30e33526265e9a3c
Tip: you can try to figure out what's wrong by changing to the process work dir and showing the script file named `.command.sh`
The nf file:
\#!/usr/bin/env nextflow
//data_location
params.outdir = './results'
params.in = "$PWD/\*.fastq"
dataset = Channel.fromPath(params.in)
params.db = "/home/blast/nt_db_20221011/nt"
process concatenate {
tag "$x"
publishDir "${params.outdir}", mode:'copy'
input:
path (x) from dataset
output:
path ("output.fastq") into cat_ch
script:
"""
cat $x > output.fastq
"""
}
process fastqconvert {
tag "$y"
publishDir "${params.outdir}", mode:'copy'
input:
path (y) from cat_ch
output:
path ("insect.fasta") into convert1_ch,convert2_ch,convert3_ch
script:
"""
seqtk seq -a $y > insect.fasta
"""
}
process blast_raw {
tag "$z"
publishDir "${params.outdir}", mode:'copy'
input:
path (z) from convert1_ch
output:
path ('blastrawreads.xml') into blastrawreads_xml_ch
script:
"""
blastn \
-query $z -db ${params.db} \
-outfmt 11 -out blastrawreads.asn \
-evalue 0.1 \
-num_alignments 1 \
blast_formatter \
-archive blastrawreads.asn \
-outfmt 5 -out blastrawreads.xml
blast_formatter \
-archive blastrawreads.asn \
-outfmt "6 qaccver saccver pident length evalue bitscore stitle" -out blastrawreads_unsort.tsv
sort -n -r -k 6 blastrawreads_unsort.tsv > blastrawreads.tsv
"""
}
I can see that the insect.fasta file has been produced and has the appropriate permissions and is located in the expected dir.
I used the following command to download the nt database
update_blastdb.pl --decompress nt --passive --source gcp
gcp is the google cloud in Australia
The nt database is ~26GiG in size.
I really need an excel, asn and fasta file from blast results for downstream analysis.
Any help would be much appreciated.

Answer 1

BLAST Database error: No alias or index file found for nucleotide
database [/home/blast/nt_db_20221011/nt]
I think you should be able to re-create the above error independently of Nextflow using:
blastdbcmd -db /home/blast/nt_db_20221011/nt -info
Note that the db argument must be a dbname, not a path. For /home/blast/nt_db_20221011/nt to work correctly, you should be able to list your db files using: ls /home/blast/nt_db_20221011/nt.*
Not sure if there's a typo in your question, but the size of the nt database is about an order of magnitude larger, at approximately 250G. I wonder if simply re-downloading the database fixes the problem? Note that you can get a list of BLAST databases (showing their sizes and dates last updated) using:
update_blastdb.pl --showall pretty --source gcp
Note also that DSL1 is now end-of-life1 and will be removed going forward. I strongly recommend migrating to using DSL2 syntax when you get a chance.
From the comments:
The problem is that when you use params to specify a path, the path or files specified will not be localized inside the process working directory when the task is run. What you want is just a second input (value) channel. For example, using DSL2 syntax:
params.db = "/home/blast/Geminiviridae_db_20221118/geminiviridae"
process blast_raw {
tag { query_fasta }
input:
path query_fasta
path db
output:
path "geminiviridae.xml"
"""
blastn \\
-query "${query_fasta}" \\
-db "${db}" \\
-max_target_seqs 10 \\
-outfmt 5 \\
-out "geminiviridae.xml"
"""
}
workflow {
db = file( params.db )
blast_raw( your_upstream_ch, db)
}