These are chat archives for dereneaton/ipyrad

1st
Nov 2016
Jenny Archibald
@jenarch
Nov 01 2016 01:43
I just sent the params file - thank you!
Alexander McKelvy
@SnakeEvolution_twitter
Nov 01 2016 15:56

I'm having trouble writing outputs in step 7. Machine has 4 available cores and 20GB of memory currently allocated. Free disk space of 700GB. The log file it points to is empty (0 bytes file size)

ipyrad [v.0.4.4]

Interactive assembly and analysis of RAD-seq data

loading Assembly: paramtest
from saved path: ~/Documents/paramtest.json
local compute node: [4 cores] on tayne2-VirtualBox

Step 7: Filter and write output files for 96 Samples
The following Samples do not appear to have been clustered in step6
(i.e., they are not in /home/tayne2/Documents/paramtest_outfiles/paramtest.hdf5).
Check for typos in Sample names, or try running step6 including the
selected samples.

Missing: 114, 130

[####################] 100% filtering loci | 0:00:32
[####################] 100% building loci/stats | 0:00:12
[####################] 100% building vcf file | 2:47:07

Encountered an error, see ./ipyrad_log.txt.
error in vcf build chunk 54600: MemoryError()

sometimes it's this
[####################] 100% filtering loci | 0:00:30
[####################] 100% building loci/stats | 0:00:11
[####################] 100% building vcf file | 1:58:46

Encountered an error, see ./ipyrad_log.txt.
error in vcf build chunk 0: IOError(Can't read data (Memory allocation failed for deflate uncompression))

I'm not exactly sure why two samples didn't cluster - there are a few failed wells on the plate. I guess help with this error and a general recommendation on at what point in the workflow to remove low quality samples would be much appreciated. My apologies if I missed either in the program documentation. Thanks!

James Clugston
@Cycadales_twitter
Nov 01 2016 19:49
@dereneaton Hi Daren, any luck with the bug in step six?
Katherine Silliman
@ksil91
Nov 01 2016 21:01
@SnakeEvolution_twitter I got the same memory allocation failure error when building a VCF on a cluster. I found the issue was I was not specifying the actual number of cores I had access to with the "-c" flag. It's possible that your computer has 4 cores, but if you were doing something else at the same time there weren't actually 4 cores available. I would suggest trying step 7 with "-c 3"
Alexander McKelvy
@SnakeEvolution_twitter
Nov 01 2016 21:04
I will try that, thanks!
Katherine Silliman
@ksil91
Nov 01 2016 21:16
@dereneaton @isaacovercast
When trying to use the VCF ouput in another program, I got this error: Error: Malformed VCF genotype '4/4:0,12,0,2', at marker '1866:25'. so I looked at the .loci file for that position and sample. It looks like whenever there is an N, ipyrad is coding it as 4/4 instead of ./. Gaps(-) are coded correctly. I'm also curious why a position would have an N, but have read depth data. Would that come from a correction during the consensus calling step? Thanks!
Deren Eaton
@dereneaton
Nov 01 2016 21:37
@ksil91 is this using the latest version of ipyrad?
@Cycadales_twitter sorry, lot's of deadlines happened today, so I didn't have a chance to work on it yet, but I do now.
James Clugston
@Cycadales_twitter
Nov 01 2016 22:42
@dereneaton thanks Daren that would be great!