These are chat archives for dereneaton/ipyrad

11th
Dec 2018
Isaac Overcast
@isaacovercast
Dec 11 2018 03:58
@tommydevitt what happened when you tried to run ipcluster by hand? What exactly did you try and what were the results? What kind of cluster is it?
jhoepke
@jhoepke
Dec 11 2018 12:57
Hello, I would like to know if it is possible to make a tetraploid snp call in ipyrad since I have GBS data including diploids and tetraploids? My goal would be to get a GT:DP:CATG-output in the vcf-file that looks like this "0/0/0/1:15:15,0,0,0" and str-files with four rows per individual. I was already thinking about to use GATK to make a tetraploid snp call but my denovo runs in ipyrad do not store any kind of bam-files whereas my reference runs do?! Many thanks in advance!
tommydevitt
@tommydevitt
Dec 11 2018 16:47

@isaacovercast I started the ipcluster instance as suggested here https://ipyrad.readthedocs.io/HPC_script.html

user@compute$ ipcluster start --n 20 --daemonize
user@compute$ sleep 60
user@compute$ ipyrad -p params-test.txt -s 2 --ipcluster

There was a pause where nothing happened, then the error message above appeared.
The cluster is a Cray XC40 (TACC Lonestar 5 https://portal.tacc.utexas.edu/user-guides/lonestar5).

tommydevitt
@tommydevitt
Dec 11 2018 17:19
@isaacovercast An update from my HPC admins: "there's a possibility that the way ipcluster starts up does in fact conflict with something related to the system. A colleague was able to test ipcontroller and found it couldn't bind to a specific socket it's attempting to do so with. It can be possible then that ipcluster is attempting to access a port that is restricted by the system. With the new changes the range of ports on a cray linux is restricted in comparison to other linux systems so there is a smaller number of usable ports. Aside from telling ipcluster which port(s) to use, there may need to be a change in the system being used (stampede2 for example)."
tle003
@tle003
Dec 11 2018 17:29
Hi, I am trying to run the 'Pedicularis' tutorial using Jupyter Notebook hosted by a cloud server. I have managed to go through the ToyTree tutorial alright for the most part; however, my problem is that I can't get sratools to download the required files off of NCBI. The error that I receive is: "AttributeError: 'module' object has not attribute 'sratools'. I look through the earlier queries here and found that someone else had problems getting their sratools to work, but they were getting a longer error message. I am wondering if this is a package compatibility or path issue, as I previously had problems getting ipyrad.analysis to run until I downloaded toytree version 0.1.6 and imported that specific version into the jupyter notebook I am working in. Currently, I am using ipa.version / ipyrad version 0.6.18 and ipyrad. Thanks - Tammy
Isaac Overcast
@isaacovercast
Dec 11 2018 18:03
@tommydevitt Well it sounds like the cluster admins are on the right track. You might need to dig into the ipcluster docs to figure out how to tune the ports its using: https://ipyparallel.readthedocs.io/en/latest/process.html
@tle003 Hi Tammy, this is a very old version of ipyrad you're using. Can you update to the newest version? I'm not sure sratools was even included in this old version, so updating will surely be more likely to work.