Problem with Haloplex variant call

[induivain]

Hi seqanwers

I am working on Haloplex data
I doing comparisons between the same sample, with two different technologies, haloplex and Sureselect, I have found the following thing In the call variants step (picture)

A great number of discrepancies are found in the last nucleotide of a lot of reads n haloplex data

I check by sanger some of these positions , and the was negative.

That I am doing badly?

I hope that you could help me or someone of agilent.

In my bioinformatic pipeline , first I remove the haloplex adaptors on the two sides.
>PrefixPE/1
AGATCGGAAGAGCACACGTCTGAACTCCAGTCAC
>PrefixPE/2
AGATCGGAAGAGCGTCGTGTAGGGAAAGAGTGTAGATCTCGGTGGTCGCCGTATCATT

with cutadapt and run in BWA-mem , and GATK best practices (don't remove duplicates)

all opinions are welcome.

Induivain