mothur

Killed: 9 while running dist.seqs


#1

Dear all,
I am running mothur v.1.41.3 on a macbook pro mid2014, 16Gb RAM and 500 Gb hard-drive (146Gb free). Files (except logfile) are stored in an external hard-drive of 2Tb (1 Tb free actually).
I was following MiSeq SOP to analysePE 16S rRNA data from 25 samples generated using bacterial primers 341f-806br.
All went well until dist.seqs that after a while stopped with message Killed: 9
I was wondering if it is a disk-space issue or RAM issue.
My string of commands are as SOP but with …

mothur > dist.seqs(fasta=/Volumes/UAB_data/0_retosMassaMare_reads/16S/0mm16Sfial.trim.contigs.good.unique.good.filter.unique.precluster.pick.pick.pick.fasta,cutoff=0.03)

Using 8 processors.

Sequence Time Num_Dists_Below_Cutoff
0 0 0
100 0 439
200 0 2333
300 1 6581
400 1 12359
500 2 19543
600 2 29085
700 3 40485
800 4 54402
900 5 69233
1000 6 87956
1100 7 107772
1200 8 129343
1300 10 152164
1400 11 178614
151500 36 574722

61200 54268 296507129
289900 54305 246614293
222700 54325 300138472
163300 54336 291592655
269400 54346 250694100
61300 54338 297459110
247100 54377 313290006
123500 54407 257338194
195400 54431 235224760
61400 54417 298379120
61500 54491 299147360
163400 56932 294298768
123600 57051 259237683
Killed: 9
MBP-mlliros2:~ marclliros2$

after nearly 24h running

Please could you give me your feedback?

cheers

Marc


#2

Hi - I suspect it died because of disc space. Did you look to see how much was left as it was running?

Pat


#3

Hi Pat
Thanks for your reply, I was looking by time to time, dist.seqs take time (it crashed after around 24-36h run), and as far as I remember it looks like there were still free space.
However, please Pat correct me if I am wrong, but resulting file from dist.seqs (whatever.dist) should be “heavy”, right?
Even crashed, there’s a file “whatever.dist” which is 220,23 Gb on disk. So I guess you’re right.
I will try running it somewhere with more memory space.

Really appreciate your comments

Cheers

Marc


closed #4

This topic was automatically closed 10 days after the last reply. New replies are no longer allowed.


#5

Hmmm, I’m not really sure what’s happening. Regardless, even if you get dist.seqs to complete, there’s no way you’ll ever be able to cluster that distance matrix. I’d encourage you to check out cluster.split and this blog post on why people get such large distances matrices: http://blog.mothur.org/2014/09/11/Why-such-a-large-distance-matrix/