Filtering spades assembly result according to coverage using SeqKit and csvtk

filter spades assembly result according to coverage.md

Filtering Spades assembly result according to coverage information in sequence header

Sample sequence

$ cat contigs.fasta 
>NODE_1_length_869844_cov_1135.34
ACTGNacgtn 
>NODE_2_length_576386_cov_975.882
acgtn

Steps

1. Converting FASTA to tabular format using SeqKit (http://bioinf.shenwei.me/seqkit/)

Note that seqkit fx2tab converts FASTA to 3-column tabular format, with sequence in the 2nd column and quality in 3rd column.

    $ seqkit fx2tab contigs.fasta 
    NODE_1_length_869844_cov_1135.34        ACTGNacgtn
    NODE_2_length_576386_cov_975.882        acgtn

2. Retrieving coverage as new column using csvtk (http://bioinf.shenwei.me/csvtk/)

    $ seqkit fx2tab contigs.fasta | csvtk mutate -H -t -f 1 -p "cov_(.+)" 
    NODE_1_length_869844_cov_1135.34        ACTGNacgtn              1135.34
    NODE_2_length_576386_cov_975.882        acgtn           975.882
   

3. Filtering by coverage (4th column) using csvtk or awk

    # seqkit fx2tab contigs.fasta | csvtk mutate -H -t -f 1 -p "cov_(.+)" | awk -F "\t" '$4>=1000'
    $ seqkit fx2tab contigs.fasta | csvtk mutate -H -t -f 1 -p "cov_(.+)" | csvtk filter2 -H -t -f "$4>=1000" 
    NODE_1_length_869844_cov_1135.34        ACTGNacgtn              1135.34
   

4. Converting tabular format back to FASTA format

    $ seqkit fx2tab contigs.fasta | csvtk mutate -H -t -f 1 -p "cov_(.+)" | awk -F "\t" '$4>=1000' | seqkit tab2fx
    >NODE_1_length_869844_cov_1135.34
    ACTGNacgtn
   

Thanks for posting this. It was quite helpful. Based on your code, I reached this solution:

seqkit fx2tab contigs.fasta | csvtk mutate -H -t -f 1 -p "cov_(.+)" | csvtk mutate -H -t -f 1 -p "length_([0-9]+)" | awk -F "\t" '$4>=10 && $5>=500' | seqkit tab2fx > filtered_contigs.fasta

for filtering assembled contigs by both coverage and length.

Just a little correction from the previous comment. It is missing an underscore in "cov_(.+)" :

seqkit fx2tab contigs.fasta | csvtk mutate -H -t -f 1 -p "cov_(.+)_" | csvtk mutate -H -t -f 1 -p "length_([0-9]+)" | awk -F "\t" '$4>=10 && $5>=500' | seqkit tab2fx > filtered_contigs.fasta