bin_refinement.sh copies non-FASTA files (e.g., .txt) into output directory, causing CheckM errors

[XREAL1TY]

Hi, thank you for developing metaWRAP — it has been incredibly helpful for my metagenomics workflow!

Bug Description

In bin_refinement.sh, the file-copying loop for input bins only filters by file size but does not filter by file extension. This causes non-FASTA files (e.g., bin.BinMembers.txt generated by some binning tools) to be copied into the output directory (e.g., binsA/), which subsequently leads to CheckM errors.

Relevant Code

https://github.com/bxlab/metaWRAP/blob/master/bin/metawrap-modules/bin_refinement.sh

for F in ${bins1}/*; do
    SIZE=$(stat -c%s "$F")
    if (( $SIZE > 50000)) && (( $SIZE < 20000000)); then 
        BASE=${F##*/}
        cp $F ${out}/binsA/${BASE%.*}.fa
    fi
done

The loop iterates over all files in the input directory. Any non-FASTA file that happens to fall within the 50KB–20MB size range will be copied and renamed with a .fa extension, corrupting the input for CheckM.

Suggested Fix

Add a file extension filter to only process FASTA files. For example:

for F in ${bins1}/*.fa ${bins1}/*.fasta; do
    [[ -e "$F" ]] || continue
    SIZE=$(stat -c%s "$F")
    if (( $SIZE > 50000 )) && (( $SIZE < 20000000 )); then 
        BASE=${F##*/}
        cp "$F" "${out}/binsA/${BASE%.*}.fa"
    else 
        echo "Skipping $F because the bin size is not between 50kb and 20Mb"
    fi
done

The same fix should also be applied to the corresponding blocks for binsB and binsC.

---

Thank you for your time, and I'd be happy to submit a pull request if that would be helpful!