These are chat archives for dereneaton/ipyrad

15th
Oct 2016
cwessinger
@cwessinger
Oct 15 2016 13:21
Hi folks, I've just switched over from pyrad to ipyrad. Any idea of what might cause the program to get hung up (~12 hours) on the chunking stage (stuck at 98%) of step 3? Thanks!
Isaac Overcast
@isaacovercast
Oct 15 2016 17:51
@cwessinger That part of step 3 should be relatively fast. Can you kill it and rerun step 3 with the -d flag? This will create a debug file called ipyrad_log.txt. If you can email me that and your params file that'd be great. iovercast@gc.cuny.edu
Edgardo M. Ortiz
@edgardomortiz
Oct 15 2016 18:09

Hello @dereneaton @isaacovercast, I have two errors happening from time to time:

  Step 7: Filter and write output files for 5 Samples
  [####################] 100%  filtering loci        | 0:00:04
  [####################] 100%  building loci/stats   | 0:00:00
  [####################] 100%  building vcf file     | 0:00:05

  Encountered an error, see ./ipyrad_log.txt.
   error in vcf build chunk 2023: ValueError(all the input arrays must have same number of dimensions)

Which I think I can avoid by excluding vcf from the output formats, but then I also had this one during step 4:

  Step 4: Joint estimation of error rate and heterozygosity
  [####################] 100%  inferring [H, E]      | 0:23:13  ERROR:ipyrad.assemble.util:  Sample PIO-Derio failed with error IndexError(index 12703 is out of bounds for axis 0 with size 12703)


  Encountered an error, see ./ipyrad_log.txt.
    Sample PIO-Derio failed step 4

The contents of ipyrad_log.txt si basically the same message:

2016-10-15 08:33:17,579     pid=44168     [jointestimate.py]    ERROR       Sample PIO-Derio failed with error IndexError(index 12703 is out of bounds for axis 0 with size 12703)
Deren Eaton
@dereneaton
Oct 15 2016 18:21
@edgardomortiz someone else reported that vcf error as well. I haven't been able to replicate it yet. The index error in step4 should be easy to fix if we can track down what's causing it. Does that sample have little data?
Edgardo M. Ortiz
@edgardomortiz
Oct 15 2016 18:30
@dereneaton, no the sample has good amount of data, I am assembling groups of five samples for d-foil. When I run the entire dataset the sample doesn't fail, however from ipyrad 0.4.4 the sample that had the most reads and clusters consistently fails step 4.