search

cmks / DAS_Tool

DAS Tool
Other
140 stars 17 forks source link

Program throws error if contig names are numbers #89

Closed mariasilviamorlino closed 1 year ago

mariasilviamorlino commented 2 years ago

Hi, I experienced an issue with DAS Tool 1.1.5 I ran the following command on an Ubuntu 18.04 system:

DAS_Tool -i DAS_tool_maxbin.txt,DAS_tool_metabat1.txt,DAS_tool_metabat2.txt -l maxbin,metabat1,metabat2 -c ../megahit_assembly_CCRD/megahit_assembly.fasta -o ./DAS_tool_out_score_01/ -t 28 --score_threshold 0.1 --write_bins --write_unbinned

Here is the output of the command:

DAS Tool 1.1.5
Analyzing assembly
Predicting genes
Annotating single copy genes using diamond
Dereplicating, aggregating, and scoring bins
Error in bmerge(i, x, leftcols, rightcols, roll, rollends, nomatch, mult,  :
  Incompatible join types: x.contig_id (integer) and i.contig_id (character)
Calls: cherry_pick -> [ -> [.data.table -> bmerge
Execution halted

Running the same command on the same system with the sample data provided in the repository gave rise to no errors.

It is useful to specify that in our lab, in metagenomics pipelines we usually rename the contigs so that they have only seuqential numbers as names:

>1
ATCCCGGAACGATCGA..
>2
GGCATATCGATGCTAGCAGTATCAGC...

We have always used version 1.1.2 with such assemblies without this problem arising. Given this and considering the error message, I suspect that the problem is contig names are parsed as integers at some point.

cmks commented 2 years ago

Interesting. Can you please send me (cmksieber@gmail.com) the log file of this run or attach it here?