Many thanks for your great works.
I try to use sub.sample() in windows10 system to resampling my .shared data. It’s a small data (15 samples, <1000 reads/sample). When I changed to centOS system with the same command, it works perfect. I’m new to this forum, and I don’t know how to attach the .shared file here. If you are interesting, I can sent the file to you by email.
By the way, there is a small question: I used the vsearch method in cluster() command to generate the *.list file, and then get the *.shared file. While I checked all the result files, and I can’t get the sequence of each OTU. I checked the get.oturep() command, and it required a *.dist file of the OTU sequences which I didn’t have based on the vsearch method. I even try to use the *.names and *.groups files instead of *.count_table file, and in the last *.shared file, I can only get the OTUs which named otu001, otu002 etc. Not my seq names.
I found that the phylotype() and make.shared() commands would get the same problem either. Could you please give me a little clue?
Sorry for my poor English.
Best wishes,
Pal
Preface: for the commands below where I had problems, I took the input files and ran them in mothur v. 1.38.1 and they ran fine.
Back to the segmentation fault question:
I had the same issue with a segmentation fault at sub-sample on Ubuntu (running on VMware workstation player) with mothur v. 1.39.5, with a large file:
sub.sample(shared=stability.an.shared, size=27208) [or sub.sample(shared=stability.an.shared), same issue either way]
Sampling 27208 from each group.
unique
Segmentation fault (core dumped)
Interestingly, no issue if I did summary.single with subsample:
(shared=stability.an.shared, calc=nseqs-coverage-sobs-invsimpson-chao-npshannon, subsample=27208)
but dist.shared totally flipped out on me if I incorporated a subsample, whether it was a specific number or “T”. Worked fine without a subsample.