qpBayes - ERROR: MCMC burn in must be less than the number of iterations

[dbssyck]

Hi Evan,

I am encountering a bug (?) when running qpbayes with a predefined number of iterations and burnin. I am attempting to run with 10 million iterations and 20% burnin but am getting the following error.

qpBayes \
    --geno filename.geno \
    --ind filename.ind \
    --snp filename.snp \
    --prefix filename \
    --pops A B C D \
    --out Out \
    --iterations 10000000 \
    --burnin 2000000  
 Traceback (most recent call last):
  File "/home/svu/dbssyck/conda_envs/qpbrute/bin/qpBayes", line 8, in <module>
    sys.exit(qpbayes())
  File "/home/svu/dbssyck/conda_envs/qpbrute/lib/python3.8/site-packages/qpbrute/qpbayes.py", line 369, in qpbayes
    calculate_bayes_factors(
  File "/home/svu/dbssyck/conda_envs/qpbrute/lib/python3.8/site-packages/qpbrute/qpbayes.py", line 280, in calculate_bayes_factors
    raise RuntimeError(
RuntimeError: ERROR: MCMC burn in must be less than the number of iterations

Strangely, when using the identical code with 9 million iterations and 20% burnin, it runs with no issues. The code is as follows.

qpBayes \
    --geno filename.geno \
    --ind filename.ind \
    --snp filename.snp \
    --prefix filename \
    --pops A B C D \
    --out Out \
    --iterations 9000000 \
    --burnin 1800000  

I tried a couple of other values above 10 million but they did not work either - is there an upper limit to qpbayes?

[ekirving]

Hi dbssyck,

Thank you for reporting this bug! The problem was that the comparison was being performed on the string values of the parameters, and not on their integer values:

>>> "2000000" >= "10000000"
True
>>> 2000000 >= 10000000
False

This has been corrected in github. To install the latest version, please run:

pip install --upgrade --no-deps --force-reinstall git+git://github.com/ekirving/qpbrute.git

[dbssyck]

Hi Evan,

Awesome, thanks for the very prompt response!