These are chat archives for dereneaton/ipyrad

Dec 2017
Dec 19 2017 03:42
@dereneaton @isaacovercast Thank you for your reply. Exactly, I owned about 600 individual samples and generated 36G of *clust files. I separated the raw reads into 44 assemblies when running step 1 to step 4, after that, I merged the 44 assemblies to a new one. Also, I removed raw fastq files and edit files to make more disk space available for afterwards steps. Now, I stopped on step 5 due to numerous temporary files, especially, I found that there were about more than 100 temporary files for one individual. The run on step 5 was stopped after 4hours with 0% consens calling. As you said the step 6 will consume more disk space than step 5, I think it will be a hard work!
Dec 19 2017 15:04

Hi @dereneaton and @isaacovercast , I have a question regarding the ABBA-BABA tests. I used the generate_tests_from_tree function and fixed p3 and p4. Our dataset contains 13 species as ingroup and two outgroup species one is relatively close to the ingroup and the other one is further out. I noticed that the outcome of the tests is quite different depending on which outgroup is selected.
Could you please describe how the loci are selected for the tests?
And what about when individuals of the same clade were pooled together for a test?