different size fastq?

[marco12345]

Hi,

I am working on bam files, and need to extract only the paired reads and output them in 2 fastq files (I'm working on a Mac's Terminal).

I extract the paired reads with samtools view:
$ samtools view -f1 -b /Input_file.bam > /pair_reads_file.bam

I then use the picard tools 1.87 version's SamToFastq to create the fastq file:
@ java -jar -Xmx2g /SamToFastq.jar I=/pair_reads_file.bam F=/Output.read1.fastq F2=/Output.read2.fastq VALIDATION_STRINGENCY=SILENT

The process creates two fastq files that should be of the same size, which I need for the following analysis. This works with most of my datas (1000 Genome data), but in some rare cases, it creates different size fastq. The number of reads in the 2 fastq files is the same, but one is bigger than the other (around 5% bigger).

Any suggestion on why is this happening?

Thank you in advance!

[Apexy]

I would imagine that if read length is not the same, file sizes would not be the same for the same number of reads. Kindly computer read lengths stats from both files.

[marco12345]

I would imagine that if read length is not the same, file sizes would not be the same for the same number of reads. Kindly computer read lengths stats from both files.

Selecting only the paired reads (samtools view -f1), the number of forward and reverse reads should be the same. In fact even in different sze fastq, the reads counts for forward and reverse is exactly the same.

[GenoMax]

Apexy was referring to the possibility that the reads in the two files may be of unequal length in places (e.g. because of QC trimming). Have you checked that?

[marco12345]

Ah, now I get it! It was actually that.
Thanks for pointing it out GenoMax, and thanks and sorry Apexy for answering and not being understood!