Diamond Clustering - Failed to allocate sufficient memory. Please refer to the manual for instructions on memory usage.

[nhngoc02]

Hi, I'm trying to run diamond clustering on a file with 9 million sequences and got this error.

I have 128 GB of RAM and 1.8T for disk storage. The database.dmnd database I used is 1.2 GB.

The command I used:

diamond cluster  -d database.dmnd -o output.tsv  --approx-id 80  --tmpdir /data  -M 64G

My log file contains:

diamond v2.1.10.164 (C) Max Planck Society for the Advancement of Science, Benjamin Buchfink, University of Tuebingen
Documentation, support and updates available at http://www.diamondsearch.org
Please cite: http://dx.doi.org/10.1038/s41592-021-01101-x Nature Methods (2021)

#CPU threads: 32
Scoring parameters: (Matrix=BLOSUM62 Lambda=0.267 K=0.041 Penalties=11/1)
Opening the input file...  [0.995s]
Input database: database.dmnd (8585030 sequences, 993623733 letters)
Temporary directory: /data
#Target sequences to report alignments for: unlimited
Database: database.dmnd (type: Diamond database, sequences: 8585030, letters: 993623733)
Block size = 12800000000
Opening the input file...  [0s]
Opening the output file...  [0s]
Seeking in database...  [0s]
Loading query sequences...  [0.638s]
Length sorting queries...  [0.601s]
Algorithm: Double-indexed
Building query histograms...  [3.384s]
Seeking in database...  [0s]
Seeking in database...  [0.004s]
Initializing temporary storage...  [0.003s]
Building reference histograms...  [1.399s]
Allocating buffers...  [0s]
Processing query block 1, reference block 1/1, shape 1/1.
Building reference seed array...  [1.48s]
Building query seed array...  [1.426s]
Computing hash join...  [0.503s]
Masking low complexity seeds...  [0.04s]
Building kmer ranking...  [0.008s]
Searching alignments...  [0.952s]
Deallocating memory...  [0s]
Deallocating buffers...  [0.004s]
Clearing query masking...  [0.069s]
Computing alignments...  [35.532s]
Deallocating reference...  [0s]
Loading reference sequences...  [0s]
Deallocating buffers...  [0s]
Deallocating queries...  [0.003s]
Closing the output file...  [0s]
Closing the database...  [0s]
Cleaning up...  [0s]
Total time = 46.054s
Reported 52104517 pairwise alignments, 52104517 HSPs.
3412693 queries aligned.
Finished search. #Edges: 97532173
Allocating buffers...  [0s]
Loading edges...  [0.391s]
Sorting edges...  [0.441s]
Computing edge counts...  [0.105s]
Computing vertex cover...  [1.334s]
Computing reassignment...  [0.104s]
Clustering round 1 complete. #Input sequences: 8585030 #Clusters: 1010961 #Letters: 115151078 Time: 48s
Temporary directory: /data
#Target sequences to report alignments for: unlimited
Database: database.dmnd (type: Diamond database, sequences: 8585030, letters: 993623733)
Block size = 3200000000
Opening the input file...  [0s]
Opening the output file...  [0s]
Seeking in database...  [0s]
Loading query sequences...  [0.268s]
Algorithm: Double-indexed
Building query histograms...  [0.287s]
Seeking in database...  [0s]
Initializing temporary storage...  [0s]
Building reference histograms...  [0.036s]
Allocating buffers...  [0s]
Processing query block 1, reference block 1/1, shape 1/1.
Building reference seed array...  [0.073s]
Building query seed array...  [0.074s]
Computing hash join...  [0.266s]
Masking low complexity seeds...  [0.07s]
Searching alignments...  [1729.5s]
Deallocating memory...  [0s]
Deallocating buffers...  [0.002s]
Clearing query masking...  [0.014s]
Computing alignments... Failed to allocate sufficient memory. Please refer to the manual for instructions on memory usage.

Could anyone help me out to resolve this issue?

Thanks in advance!

[bbuchfink]

There are some issues causing increased memory use that will be fixed in the next release. For now one thing you could try is using --bin 256 (or possibly higher).

[bbuchfink]

Another option would be --cluster-steps faster_lin fast_lin, that should be sufficient for 80% id cutoff.