SAMTools showing incorrect depth for a given position

[rdeborja]

I have a bam file and output from both samtools 0.1.12a and bedtools 2.11.2. The depth provided by samtools pileup is ~4000, samtools mpileup is 592 and bedtools coveragebed 12,995. The bedtools value has been confirmed when performing a samtools view for the given chromosome and position of interest.

Has anyone else encountered a descrepancy between these values as described above. Someone has mentioned something similar on the Samtools mailing list but no responses have been provided.

Cheers,
Richard

[rdeborja]

Looks like it was a simple solution. Needed to upgrade to the latest version of Samtools (we were using 0.1.12) . Seems like that version of Samtools will use only properly paired reads in the BAM file. If you you use 0.1.13 or later, Samtools mpileup has a -A flag which will use anomalous read pairs in SNP/Indel calling.