Prinseq-lite problem

[kumar03]

Dear all,
I am using prinseq (v.0.20.4, CentOS) to filter my paried end data with the following commands.
[root@localhost]# perl prinseq-lite.pl -fastq S1_read1.fq -trim_right 1 -trim_qual_right 10 -min_len 100 -derep 123 -out_format 3 -out_good S1_read1.trim -out_bad null -log
Estimate size of input data for status report (this might take a while for large files)
done
Parse and process input data
done
Check for duplicates
done
Write results to output file(s)
done
Clean up empty files
done
Input and filter stats:
Input sequences: 489,463
Input bases: 49,435,763
Input mean length: 101.00
Good sequences: 292,060 (59.67%)
Good bases: 29,206,000
Good mean length: 100.00
Bad sequences: 197,403 (40.33%)
Bad bases: 19,937,703
Bad mean length: 101.00
Sequences filtered by specified parameters:
derep: 197403

Output: log file and S1_read.trim.fastq

[root@localhost]# perl prinseq-lite.pl -fastq S1_read2.fq -trim_right 1 -trim_qual_right 10 -min_len 100 -derep 123 -out_format 3 -out_good S1_read2.trim -out_bad null -log
Estimate size of input data for status report (this might take a while for large files)
done
Parse and process input data
done
Clean up empty files
Use of uninitialized value $fhmappings in unlink at ../prinseq-lite-0.20.4/prinseq-lite.pl line 1833.
done
Input and filter stats:
Input sequences: 489,463
Input bases: 46,498,985
Input mean length: 95.00
Good sequences: 0 (0.00%)
Bad sequences: 489,463 (100.00%)
Bad bases: 46,498,985
Bad mean length: 95.00
Sequences filtered by specified parameters:
min_len: 489463

Output: only log file
Could anybody let me know what the problem is and how to solve it?

[Brian Bushnell]

You would be much better off using a program that can handle paired reads, like BBDuk. Prinseq is obsolete, IMO.

[kumar03]

There is no output based on -min_len 100. Here my PE file one consist of 101 bp and another one 95 bp by checking fastqc. May be this reason no output was generated. In this case how to assign minimum length ?
note: Initially illumina machine produced an average of 101 bp per samples by Paired-end seq.

[SES]

Prinseq handles paired-end files, it is not obsolete. You need to specify them correctly on the command line though with '-fastq' and '-fastq2' for each pair. It generates beautiful graphs of the data which gives you a much deeper understanding of the results than anything else. Most trimming programs either fail or succeed silently which is useless for quality control and leads to a lot of wasted time trying to understand the output.

Yes, I think setting the minlength to 100 is too high, especially if your reads are 95bp . I would set it lower, like 40, so you do not discard too many good reads by length following quality trimming.

[kumar03]

Thanks for your reply.