Metaspades stalling on DistanceEstimator at K55

[katebowie]

Description of bug

Hi! I've been working with the anvio-8 snakemake workflow and using metaspades within that. I was trying out the workflow on 3 samples. The first sample finished the whole process completely, while the second sample has stalled out at the "DistanceEstimator" step at K55. The log won't update for days until I cancel the job.

I read that the conda installed spades could be a little funky, so I installed the github version 4.1.0 of spades and replaced that in my conda environment. I then have re-run just on the struggling sample, and I'm still having the issue of stalling in that exact same spot. I've also checked the quality of my sample and it is similar to the sample that worked fine.

Thanks for your time!

spades.log

spades.log

params.txt

params.txt

SPAdes version

SPAdes v4.1.0

Operating System

OS: Linux-4.18.0-477.74.1.el8_8.x86_64-x86_64-with-glibc2.28

Python Version

3.10.14

Method of SPAdes installation

conda, then binaries

No errors reported in spades.log

• [x] Yes

[asl]

@katebowie This is pretty strange, the process is pretty deterministic there. Is there by any chance a possibility for to share data with us so we can reproduce?

[katebowie]

@asl Sure, happy to. What is the best way for me to share the data with you? I can share the sample that worked and the one that keeps stalling out.

[asl]

We just need the one that stalled :) Working one is, well, working. We are pretty flexible with data sharing. And you know the size and limitations on your side.

[katebowie]

Here is a link to a google drive with gzipped files for this sample: https://drive.google.com/drive/folders/1uKCk4CnsmUbGt5mzrpeeiOJUvgNJNBlJ?usp=drive_link

[katebowie]

@asl just checking in - were you able to get the data ok?

[rubenluz]

I have encountered the same problem with several different samples and systems. I get stuck at this step for hours, but I am always able to finish, but after several hours only. So same amount of data and different samples that take 1 hour normally, can go up to 12/24 hours stuck at that step.

@katebowie @asl any solution?

[aforestsomewhere]

Chiming in that I am also seeing this on certain samples, on my conda install.