My fastq files(paired end) have different read number

Question

[kimgeng]

Hello, I'm doing mapping & Calling of my data, but, I have some question about my data.

Here is the result of "samtools flagstats" of my bam file (created by bwa aln)

but, I fount that two files have different read counts. although it is paired-end data.

I believed that paired-end sequencing data must have same amount of reads, but, my data does not. And, as you can see that there is a lot of singletons and very small proportion of "properly paired"

Does my data have problem? Should I remove my data and try again..?

[colindaven]

Only 2.9% are properly paired. This is bad. You could tell use a bit more about your sequencing project and application and especially ref used so we can help.

Answer 1

[ntsopoul]

Yes, this is normal because depending on how you generate the bam file, the non-aligning reads are saved in the .bam file along the properly aligned.

[kimgeng]

Oh, I see, Thank you however, I want to make vcf files with gatk haplotypecaller and GenotypeGVCFs. But I got some problems I have about 2.6 10^8 mapped reads numbers, however, as you see, I have only 7.9 10^6 properly paired read counts.

maybe, ProperlyPairedReadFilter of HaplotypeCaller can filter and call only properly paired reads, but, in that case, I have very small amount of variants. So I wonder which one is more proper. (call only properly paired / call all of the paired mapped reads)