samtools mpileup and duplicate reads

[NestorNotabilis]

Hi,

By default, samtools mpileup ignores reads flagged as duplicates (i.e., marked with bit flag 1024 in the BAM). Does anyone know of a way of getting samtools to include duplicates in the mpileup output without having to strip the 1024 flag from the original BAM first?

Many thanks.

[Richard Finney]

Looks like one way is to change the code for a "one-off" special purpose mission. Make a new copy of samtools. Keep it separate from the real samtools.

Do this ....
Edit bam_plp_init() function in bam_pileup.c

bam_plp_t bam_plp_init(bam_plp_auto_f func, void *data)
{
bam_plp_t iter;
iter = calloc(1, sizeof(struct __bam_plp_t));
iter->mp = mp_init();
iter->head = iter->tail = mp_alloc(iter->mp);
iter->dummy = mp_alloc(iter->mp);
iter->max_tid = iter->max_pos = -1;
iter->flag_mask = BAM_DEF_MASK;
iter->flag_mask &= ~BAM_FDUP; // ADD THIS LINE turn off dupe flag

make and your ready. Just point to the right version.

[NestorNotabilis]

That's brilliant Richard, exactly what I needed. Will give it a go. Cheers!