HUMAnN3 Diamond Cannot Find tmp file

Hi, I am trying to run my fastq files with humann3. But I’m getting a critical error message when humann3 is attempting to run diamond. I have diamond 0.9.36 installed. I’m not sure why this is happening, could you help me? Thank you!

The error message:

Running diamond …

Aligning to reference database: uniref90_201901b_ec_filtered.dmnd

CRITICAL ERROR: Can not find file /oasis/tscc/scratch/yiz023/A033_10_fastq/A033_10_T0_merged_humann_temp/tmp2g66y6/diamond_m8_5Lv3R4

The command I’m using:

humann --input A033_10_T0_merged.fastq --output A033_10_fastq

The databases I downloaded:

humann_databases --download chocophlan full humann_dbs
Download URL: http://huttenhower.sph.harvard.edu/humann_data/chocophlan/full_chocophlan.v296_201901b.tar.gz
Downloading file of size: 16.01 GB

humann_databases --download uniref uniref90_ec_filtered_diamond humann_dbs
Download URL: http://huttenhower.sph.harvard.edu/humann_data/uniprot/uniref_ec_filtered/uniref90_ec_filtered_201901b_subset.tar.gz
Downloading file of size: 1.61 GB

UPDATE:
I tried again by running the demo file. This time it gave me another error saying cannot find executable diamond, but I do have diamond installed.

Decompressing gzipped file …

Running metaphlan …

Found g__Bacteroides.s__Bacteroides_dorei : 57.96% of mapped reads
Found g__Bacteroides.s__Bacteroides_vulgatus : 42.04% of mapped reads

Total species selected from prescreen: 2

Selected species explain 100.00% of predicted community composition

Creating custom ChocoPhlAn database …

Running bowtie2-build …

Running bowtie2 …

Total bugs from nucleotide alignment: 2
g__Bacteroides.s__Bacteroides_vulgatus: 1195 hits
g__Bacteroides.s__Bacteroides_dorei: 1260 hits

Total gene families from nucleotide alignment: 545

Unaligned reads after nucleotide alignment: 88.3095238095 %

Running diamond …

Aligning to reference database: uniref90_demo_prots_v201901b.dmnd

CRITICAL ERROR: Can not find executable diamond

I think there are a couple of different issues at play here. In the first case it looks like the HUMAnN task is having trouble creating one of the temp files it created during the translated search step. Were you about to confirm that this file exists after the task terminated? Seeing the scratch in the file path, I wonder if it might’ve been auto-cleaned before the task completed?

In the second example, it looks like you’re running in a different environment (or perhaps didn’t have all the modules loaded that you were using in the first environment?). You should be able to check if DIAMOND is available by running diamond --version.