pontussk/popstats

Population genetic summary statistics

popstats's Issues

invalid literal for int() with base 10: 'C'

Hey,

I am have converted my vcf file using the vcf2tped.py script. After that, I want to try to estimate FAB. But when I try to run it using the command:
python popstats.py --file input.tped --pops INDV1,INDV2,INDV3,INDV4 --ancestor ANC_INDV --FAB

I get the error:
Traceback (most recent call last):
File "/home/uramakri/ryanr/softwares/popstats/popstats.py", line 710, in
chromosome = int(col[0].lstrip('chr'))
ValueError: invalid literal for int() with base 10: 'CM000001.4'

Could you please let me know how to resolve this? the reference genome is dog not humans.

Anubhab

f3vanilla requests 4 populations

Hey Pontus!

I tried running popstats with the --f3vanilla option and it seems to require four populations to be specified, even though the --f3 correctly asks for 3.

$ python2.7 popstats.py --f3vanilla --pops 1,2,3 --file F3_popstats.merged 
Traceback (most recent call last):
  File "/Users/lamnidis/Software/popstats/popstats.py", line 185, in <module>
    poplabel4=poplist[3].split('+')
IndexError: list index out of range


$ python2.7 popstats.py --f3vanilla --pops 1,2,3,3 --file F3_popstats.merged 
#Runs without raising an error.

while --f3 runs normally with 3 populations:

$python2.7 popstats.py --f3 --pops 1,2,3 --file F3_popstats.merged 
#Runs without raising an error

Thank you!
Thiseas

Defining pops in tfam file

Hi Ponstuck,

I have a tped file with all autosomes and tfam file with the individuals (286) from 7 different populations. Each of the pops have a specific identifier the consists of a 3 letter pop code and individual identifier e.g KPA_562. I was inquiring how I would use the python script to extract the pops I want to study? The example file in the link you had provided in one of the issues I read is not active.

Thank you.

--pi not an option

Hi,
I recently started using your code to estimate f2, since I couldnt seem to get AdmixTools to compile correctly. I have got most of the tests listed in the README to work correctly, although --pi doesnt seem to be recognized as an option. I also cant find it listed in the python file as an option. Any plans to update this?

Also I couldnt find the vc2tped.py mentioned in the README so just used VCFTOOLS --plink-tped option.

Any chance you might implement something where --pop option could accept a "file" with a list of which individuals belong to which populations? Currently, it is a bit cumbersome to list all individuals that belong to a population on the command line (e.g., ind1+ind2+ind3,ind4+ind5+ind6) and for calculating f2 needs to be rearranged since it only looks at pops in 1 and 2 position. What about a pairwise calculation for anything involving only 2 pops?

thank you,

About the input files

Hi, I am not clear about the input files. For example, I have Pop1, Pop2, Pop3 and Outgroup. I want to calculate D-statistics (Pop1 , Pop2 ; Pop3 , Outgroup). Do i need to put these four populations in one tped file? or in 4 tped files ? Thanks.

Basic question of input files

Hi again Pontusk,

I was wondering if I use an input that has not not been filtered for Bi-allelic SNPs only will popstats just automatically ignore those SNPs with more than two alleles or will I have to provide an input that is already filtered for bi-allelic SNPs only?

Best,
Alex

DArTseq Data to popstats

Hello,
I am using DArTseq data in my analysis. I have missing values and this data type is binary. I was able to upload my data file to Plink and then I transformed them to use in popstats. However, I get errors when I tried to upload it to popstats. Do you think this is due to my data type and missing values?
Thank you :)

Best regards
Buddhika

Question about 'Dp_main = sum(t_list) / sum(n_list)'

Hi Pontus
I am trying to run the D-statistics. I followed your suggestion to put the population name in the first column. Then I feed popstats.py with my .tped and .tfam file then it shows

########################################
Traceback (most recent call last):
File "popstats.py", line 2105, in
Dp_main = sum(t_list) / sum(n_list)
ZeroDivisionError: integer division or modulo by zero
########################################

Command I used was
python2 popstats.py -p good.tped -f good.tfam --not23
-b 1000000 --pops pop1,pop2,pop3,pop4 --informative

What kind of error in my data lead to this?

Thank you!

sample larger than population

Hi,

I get a sample size error when I try to execute:

python popstats.py --file input --pops POP_CI,POP_SI,POP_AR,POP_TH --ancestor POP_ANC --FAB

I have only one ancestral individual. Does the program need more than one outgroup? do all outgroup individuals have to be from the same speceis?

Thanks
Anubhab

about h4

Dear Pontus,

I was trying to understant the parameters for the h4 test (LD option) and I have some doubts

the --LD option is the classic D of LD but what about the LD window? which are the unit for this option? I thought that were cM but the default is 5000 (in the code).

Thank you for your time