hi folks,

I'm running cd-hit on a fasta file containing 8.7 million fasta sequences. I've taken chunks from this file (up to 400,000 sequences so far) and run on cd-hit, with no problems. However, when I run the full file (nohup ./cd-hit1 -i test.fasta -o test-out-cluster -c 1.0 -n 5 -M 2000&), it takes a day or so, then the job ends, and the output cluster file remains empty. The nohup.out file ends with

..........3712000 finished 3231208 clusters
.SEG 2 3712119 7064775
Reading swap
Comparing with SEG 1

Now, I wonder if there might be another error message than nohup isn't picking up on.

I'm working on a server with this memory.

total used free shared buffers cached
Mem: 16380068 10612808 5767260 0 107004 8076380
-/+ buffers/cache: 2429424 13950644
Swap: 10241364 140 10241224

There's two possibilities I can see:
1. my server is running out of memory in some way
2. cd-hit doesn't like dealing with such large input files (8.7 million seqs).

Could anyone confirm that cd-hit has been used with a file of similar size before? This would allow me to discount the second possibility.

best regards, and thanks for any help.

Paul