VCFtools: --window-pi and --from-bp --to-bp produces different start and ending windows

Question

[GabrielMontenegro]

I am trying to estimate pi on a given region, therefore I used the --from-bp and --to-bp functions to specify it:

THE LOG FILE

Parameters as interpreted:
--vcf file
--chr 5
--to-bp 33976693
--out test
--window-pi 50000
--from-bp 33926693

After filtering, kept 103 out of 103 Individuals
Outputting Windowed Nucleotide Diversity Statistics...
After filtering, kept 1482 out of a possible 5033226 Sites
Run Time = 33.00 seconds

The region specified has exactly 50000 bp and therefore I use the --window-pi 50000 option to get the nucleotide diversity of that whole region, BUT

RESULT

CHROM BIN_START BIN_END N_VARIANTS PI
5 33900001 33950000 108 0.000323596
5 33950001 34000000 113 0.000431332

My region is within those two boundaries but this is not what I wanted... Also the N_VARIANTS doesn't add up to the 1482 kept variants that the log file is reporting.

[amitgourav.ghosh12]

Hello Sir, I am facing the same problem. I somewhat managed to get one whole region to come as one in Tajima's D calculation but I am unable to do so for the pi calculation.

Did you find a way to overcome this problem? I yes, can you please suggest me what to do?

Thanking you, Amit