These are chat archives for dereneaton/ipyrad

21st
Jun 2017
Deren Eaton
@dereneaton
Jun 21 2017 17:10
@toczydlowski and others, there is now support for treemix (and a few other new outputs) through the ipyrad API (i.e., you run a bit of code in a Python session after you finish your assembly). See here for examples: http://ipyrad.readthedocs.io/analysis.html
toczydlowski
@toczydlowski
Jun 21 2017 18:19
@dereneaton Great! Looking forward to trying it out. Thanks.
Nic Diaz
@NebulousNic_twitter
Jun 21 2017 20:41
image.png
@dereneaton @isaacovercast Hi again, after making it through ipyrad we decided to use your jupyter cookbook for TreeMix. After we set everything up, all cells seem to be running until the cell: #you can run the command in a notebook by using bash one-liners, where we get the following error:
Deren Eaton
@dereneaton
Jun 21 2017 20:43
@NebulousNic_twitter Hmm, I haven't added any safety checks into the code yet to ensure that the data matrix isn't empty, so my guess is that maybe that is the problem. Take a quick look at the input file .treemix.in.gz and make sure there is at least some data in it. If not, then trying decreasing the minmap values, or doublecheck the spelling of the sample names.
Nic Diaz
@NebulousNic_twitter
Jun 21 2017 21:10
@dereneaton the input file seems to only have the 'species' names I entered into the dictionary. minmap values were all set to 1.
Deren Eaton
@dereneaton
Jun 21 2017 21:32
@NebulousNic_twitter , my guess is that there may be no SNPs in your dataset where there is no missing data in at least one or your populations. You could try including fewer populations at a time, or re-assembling your data set with a lower mindepth and minsamp settings to increase the amount of shared data.
The next update to the treemix object will report the number of SNPs that pass filtering.
leclearnm
@leclearnm
Jun 21 2017 22:51
@dereneaton I am calculating the site specific likelihood score for a outfile.phy concatenated matrix in RAxML using -f G given two conflicting topologies, and I would like to be able to sum the individual scores within a given locus into a single value in order to identify genes that are strongly supporting alternate topologies (similar to Shen et al., 2017 Nature Ecology and Evolution). Does any output from ipyrad contain position information for loci boundaries in the .phy file, (like if you were supplying partitions to mrbayes)?
Deren Eaton
@dereneaton
Jun 21 2017 22:58
@leclearnm we only write partition info for the .snps file at the moment, but we should produce it for the full phylip file as well. We will probably write the partition info directly into the nexus output (which is something pyrad does). I raised a ticket for it #249
leclearnm
@leclearnm
Jun 21 2017 22:59
@dereneaton Great! Thanks. Right now I am working with the snps.map file to find a work around.
Meanwhile, I am still having issues with the ifc.loci2multinex command. It is only writing, at most, 80 individual nexus files from .loci file for my largest min4taxa matrix, which has >7500 loci. That is with no taxa subsetting in the bucky.ipynb