ablab
metaSpades crashes with merged reads
Description of bug
Hi,
I run metaSpades on a large set of samples. It works fine when I just use the pair1 and pair2 files. But I also have merged reads for each library that I would like to use. I added them to the yaml input file under "merged reads". But when I run spades then, it first runs for a while and then crashes with the following error:
0:07:47.455 0M / 4247M INFO General (read_converter.cpp : 92) Converting merged reads
spades-core: /spades/assembler/src/common/io/reads/async_read_stream.hpp:101: void io::AsyncReadStream<ReadType>::init() [with ReadType = io::SingleRead]: Assertion 'is_open()' failed.
== Error == system call for: "['/data/vanessa/programs/SPAdes-3.15.4-Linux/bin/spades-core', '/data/bigexpansion/vanessa/ragweed/2022-07-01_metagenomicAssembly/2022-09-05_ragweed_SPADES_modern_noBGI_pair1_pair2_merged/K21/configs/config.info', '/data/bigexpansion/vanessa/ragweed/2022-07-01_metagenomicAssembly/2022-09-05_ragweed_SPADES_modern_noBGI_pair1_pair2_merged/K21/configs/mda_mode.info', '/data/bigexpansion/vanessa/ragweed/2022-07-01_metagenomicAssembly/2022-09-05_ragweed_SPADES_modern_noBGI_pair1_pair2_merged/K21/configs/meta_mode.info']" finished abnormally, OS return value: -6
None
======= SPAdes pipeline finished abnormally and WITH WARNINGS!
I tried to increase memory, but that doesn't help. It crashes at exactly the same point again. Am I doing something wrong when adding the merged reads to my yaml file? Or is it not possible to use merged reads with metaSpades?
Vanessa
spades.log
params.txt
SPAdes version
SPAdes 3.15.4
Operating System
Linux-4.15.0-156-generic-x86_64-with-debian-buster-sid
Python Version
Python 3.6.10
Method of SPAdes installation
binaries
No errors reported in spades.log
- [X] Yes
Will you please attach your /data/bigexpansion/vanessa/ragweed/2022-07-01_metagenomicAssembly/2022-09-05_ragweed_SPADES_modern_noBGI_pair1_pair2_merged.yaml file?
Here is the yaml file. Sorry that I forgot to attach it before
2022-09-05_ragweed_SPADES_modern_noBGI_pair1_pair2_merged.yaml.zip
It might happen that you hit the # of open files limit. Can you concatenate (in the same order of files) you left reads, your right reads and your merged reads – does the problem reproduce then?
Thank you for your fast reply. I have now tried to run it with a subset of samples and it finished without errors. The the number of files seems to indeed be the problem.
To make sure I understand right: with concatenating you mean to create one file containing all the pair1 reads, one file for all the pair2 reads and one file for all the merged reads, right?
Not all of my sample have the same read length (some are 100 bp, most are 150 bp). Would that be a problem when I concatenate them all together? So would you recommend to split them based on read length?
To make sure I understand right: with concatenating you mean to create one file containing all the pair1 reads, one file for all the pair2 reads and one file for all the merged reads, right?
Correct
Not all of my sample have the same read length (some are 100 bp, most are 150 bp). Would that be a problem when I concatenate them all together? So would you recommend to split them based on read length?
Not a problem at all