BWA alignment and very low properly paired

Question

[seta]

Hi all,

I would like to see HLA allele at the BAM file level using IGV as HLA sequences as reference in IGV. So, using samtools, I extracted the HLA region from bam file resulting from aligning the fastq reads (from whole genome sequencing) with hg38 and after converting to fastq reads by SamTofastq (from Picard), I aligned the fastq reads to HLA genomic sequences using bwa mem -L 10000 -a; although the mapping percentage looks good enough (94%), the properly paired is very low (about 1%)! Could you please share me your idea why the properly paired is so low?

I also run BWA mem with default parameters, which returned me just about 20% mapping percentage

Also, based on samtools flagstat of extracted BAM file of HLA region, the properly paired is well (about 98%).

Please kindly share me your idea.

Thanks

[swbarnes2]

Are you absolutely sure that the fastqs you made with Picard are properly paired? Did you check read names?

[seta]

As I checked with head and tail, the read names in both fastq file are the same. What else shall I check to solve the issue?

Also, during mapping, I frequently got the warning that "skip orientation FF as there are not enough pairs". The same warning also appeared for FR, RF, and RR. However, as I mentioned in the post, there is more than 90% properly paired in the extracted bam file based on samtools flagstat.

Any suggestions and comments would be highly appreciated.