Treemix - wrong values of m and i in OptM

[Zoe Meziere]

Hi all,

I am using the OptM R package to determine the optimal number of migration edged from my Treemix runs. However, when using the function optM(), it recognises the wrong number of migration edges and iterations. In my bash script (pasted bellow), I used tm varying from 1 to 10 and using 10 iterations, but optM() outputs "m ranges between 0 and 4. The average number of iterations per m (not including m = 0) was 25.". The Treemix output files are correctly labelled so I don't know where the issue is coming from...

Have you had this issue before?

Cheers,

for m in {1..10}

   do

   for i in {1..10}

      do

      s=$RANDOM

      echo "Random seed = ${s}"

      treemix \

         -i Stylo_treemix_input.frq.gz \

         -o Stylo.${i}.${m} \

         -global \

         -m ${m} \

         -k 500 \

         -seed ${s}

      done 

done

[Haley Arnold]

Hi, I am experiencing the same issue. Did you ever figure out what was going on? Is this normal optM behavior?

Thank you,

Haley

[Dylan O'Hearn]

I believe I've figured this out.  When it reports the range of m values and how many iterations there were of each, it's basing that not on how many migration edges you allowed in the runs (1-10) but how many were actually used in each run, since Treemix is evaluating all values up to the number you allow and then going with whichever has the highest likelihood.  So when it says m ranged from 0 to 4, that means that none of the runs resulted in an m value >4 having the highest likelihood. So I don't think it's actually a problem although it is very confusing.