Dear all, my RAXML run was interrupted after all bootstraps were completed. The RAXML specialists advice: "However, since you already have the bootstraps, you can now independently compute a couple of ML trees (see manual) and then use these results to do post-analyses ..." (https://groups.google.com/forum/?hl=en#!searchin/raxml/interrupted$20run%7Csort:date/raxml/BoFLQYrBlT8/pqLRqIgWxVYJ)
Unfortunately I cannot find the needed command in the manual. Can anybody, please, point me the right place or explain how to build the best-scoring ML tree and continue with the analysis if the raxml bootstrap file has been generated already?
thank you for the help in advance! :)
•
written
by
victoria_aleks •
1 answer
By default, RAxML doesn't perform bootstraps. If you already have the bootstraps but didn't perform the ML searches for best tree, first run RAxML with -f d(whichis the default) or -f o (slower, but slightly more accurate likelihood scores), then use -f b to draw bootstraps onto the best tree.
So for example, to perform 100 searches for best trees, each search starting from a different random tree:
raxmlHPC -f d -m GTRCAT -p 12345 -s alg -n MLTrees -N 100
Then, to add the bootstraps to the best tree:
raxmlHPC -f b -t bestTree -z bootstrapTrees -m GTRCAT -n best+boot
Make sure to pass different random number seeds to RAxML and not
only 12345 as I have done in the examples.
Log in to answer this question.