When I use .arp files in arlequin to run mismatch distribution it says that can’t converge beyond 2000 steps and also no graphs are produced.can someone suggests what can I do to overcome this problem

  • $\begingroup$ Hi, which analysis are you performing? $\endgroup$ – M__ May 8 '20 at 8:28
  • $\begingroup$ it is on mitochondrial CO1 DNA sequences and I am performing mismatch distribution using .arp file $\endgroup$ – Upasana Singh May 8 '20 at 11:25
  • $\begingroup$ How many sequences did you supply? $\endgroup$ – M__ May 8 '20 at 11:46
  • $\begingroup$ 59 DNA sequences of length 563bp, but it is a haplotype file and it has 9 haplotypes within $\endgroup$ – Upasana Singh May 8 '20 at 12:46

What you appear to have is a Bayesian MCMC for 2000 "steps" which isn't a lot and normally the term 'steps' is used for parsimony, so I'm a bit confused.

However, if a Bayesian analysis does not reach convergence, i.e. duplicate runs do not achieve the same results then the data space has not been solved so you get no result.

9 haplotypes does not sound much to me, it is possible there are too many sequences, it is also possible there is simply not enough variation (9 haplotypes ... how many alleles?).


  • Arlequin should provide a default convergence value which you can lower to force a result
  • Decrease the number of sequences but maintain the number of haplotypes
  • Try their example data set

Do you get any associated file such as posterior probability, any output file with a minus sign in front of it?


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.