Hello guys, I'm testing the SMS to select best substitution models on alignment (http://www.atgc-montpellier.fr/phyml-sms/), but, due the large number of sequences in my alignments (6413) I'm testing in a local server.
My problem is that this tool appear accept only datasets with 4000 or less sequences, and when I tried to run with my dataset the error of 4k sequences occurred.
Starting SMS v1.8.4
Input alignment: teste.phy
Data type: DNA
Number of taxa: 6413
Number of sites: 228 Number of branches : 12823
Criterion : AICStep 1 : Set a fixed topology Unexpected error while computing Log-likelihood
The number of taxa should not exceed 4000phyml: io.c:1293: Get_Seq_Phylip: Assertion `FALSE' failed.
Someone here know the best way to avoid this situation? (with SMS or with other tool to select the best model of nucleotide or aminoacids in large scale?).
Best,
Ps: I run this tool with a subset of 3k sequences, and everything runs ok...
1 answer
After an e-mail to author of SMS, this is how can I've solved this problem:
I download the source code of phyml on https://github.com/stephaneguindon/phyml and I modify the value of variable N_MAX_OTU from /src/utilities.h. After this, I compiled the phyml, and copy all files of phyml directory to sms/phyml-src/phyml-3.3.20190909/.
So these steps allow the use of SMS with alignments with more than 4,000 sequences....
Log in to answer this question.