Hi all,
I am running a single fastq file for HUMAnN in cluster with the following code:
conda activate humann
humann -i file.fq -o file_humann --threads 40
I am running this command on a cluster with 24CPUs and 50Gb memory with 40 threads.
I am getting the ERROR: Unexpected end of input files.
Could anyone please guide me in this? The error screenshot is pasted below-
I tried running demo fastq file provided with HUMAnN, yet I am getting the similar error, might be something to do with my HUMAnN installation. Any leads? Error is thrown from Diamond
Hmm, my first guess was the number of threads - we have seen errors in alignment with very large numbers of threads getting out of step with file I/O, leading to errors like this. In your demo run though it looks like you’re only using one thread though.
My next guess is an incompatibility with the newest version of DIAMOND (2.0.15). HUMAnN 3.0 was notably built/tested against DIAMOND 0.9.36, the end of the pre-2.0 lineage. I have seen it used successfully on DIAMOND 2.0+ but it’s possible something has changed recently that resulted in this error. I would try rolling back to an earlier DIAMOND and see if that fixes the problem.
I recently face the same Error, when Humann tested and ran the Demo fastq file and got the same message, then I notice the message before CRITICAL ERROR and the Error message returned:
Running diamond …
Aligning to reference database: uniref50_201901b_ec_filtered.dmnd
That diamond align to the first database and failed, maybe I need to update it, so I tried to update the diamond database and rerun the demo, finally, I run the entire process and got the demo result without error.
I run this command on a WSL2 with 12 CPUs and 24Gb memory with 20 threads.