These are chat archives for dereneaton/ipyrad

6th
Jun 2017
Jenny Archibald
@jenarch
Jun 06 2017 14:23 UTC
@isaacovercast @dereneaton Thanks for all your help so far - ipyrad finished its analysis on my data! Now I am wondering about tetrad. I started an analysis about 18 hrs ago, and it is still showing 0% progress. I know the website warned it was slow with lots of accessions (I have 288), and so maybe it's just that slow, but I wanted to check in to see if it is likely to progress or if something is wrong. This is what the output currently shows:
-------------------------------------------------------
tetrad [v.0.6.19] (ipyrad.analysis toolkit)
Quartet inference from phylogenetic invariants
-------------------------------------------------------
loading seq array [288 taxa x 2513283 bp]
max unlinked SNPs per quartet (nloci): 220722
tetrad instance: ch4tet
host compute node: [16 cores] on m027
inferring 280720440 induced quartet trees
[                    ]   0%  initial tree | 18:27:50 |
Isaac Overcast
@isaacovercast
Jun 06 2017 16:37 UTC
@SheaML Yeah, those happen every once in a while, in rare circumstances R1 or R2 fail to map correctly so the vsearch merge fails. It's very infrequent so you can just ignore it.
Shea Lambert
@SheaML
Jun 06 2017 16:59 UTC
@isaacovercast Thanks. I do have 200,535 lines in my ipyrad_log.txt file, formed in the ~20hours I've been mapping samples (all errors of this type), in case that sounds like an unusual number to you. Also, I've been stuck on the "finalize mapping" step for 18+ hours now (0% progress).
Isaac Overcast
@isaacovercast
Jun 06 2017 18:17 UTC
@SheaML What version are you running?
Deren Eaton
@dereneaton
Jun 06 2017 18:45 UTC
@jenarch Hi Jenny, as you can see, the number of possible quartet trees for a 288 taxon tree is huge (280M), and so it will take a pretty long time to run. For a tree of this size you will probably not want to sample all possible quartets, but instead randomly subsample just a subset of quartets. Maybe try sampling around 10M or so to make it run in a reasonable amount of time. And however many cores you can get the better. The speed should increase approximately linearly with more cores.
Shea Lambert
@SheaML
Jun 06 2017 19:20 UTC
@isaacovercast 0.6.27. Steps 1 and 2 were run on 0.6.25, though.
Jason Cantley
@jaecan808_twitter
Jun 06 2017 22:29 UTC
Hi all, I'm running into an early error using ipyrad. I can create an assembly, but but early on it appears the ## [2] parameter is failing recognizing my fastq.gz file. The error we get reads as follows "ERROR Not a gzipped file". Can anyone point me in the right direction? Many thanks! Jason