replacing a part of row name by sample name

Question

I have multiple samples named p1_500sc.fasta, p2_500sc.fasta... Each file loks like this:

>NODE_60_length_504_cov_1.580902
TATATCGCCGTAGATAGACGAATACGGGATTTTGAAATGACTGATATATTCAAGCACATC
CTGTTCAGTCCAAAATGACATCGGCTTCGATTTCGGGATTGCCCCCGCAAAGTTGTTACA
>NODE_61_length_503_cov_4.457447
GATATGGGCCGGTGAGCATGCTGTCACATTTTGGGCAGGTGCCGAGGATTATGAGCTCGT
CTTCCGGCGTCAACGCTTGTTCGTTGCGTCTGGTGATGTGTTCCAGAGCGGCATAGTCGT
>NODE_62_length_500_cov_4.037534
CTCGTTCTTGCACTCCTGAATGAAGCGAATCTCTGACGAGGGTATTCCATAATCCTCTAT
CAGCTTACGCTTGATTTCTG
>NODE_63_length_500_cov_1.718499
GGAGGATATCAATGCCCTTATACTGCTTCTCAACAGTAGGAGTAATGCGGATACCCAAAA
CGGTCTTAGGGTTCTCCAGCAACTTCATGGCATTCCAACGCTTCAAGTCATCCATGCGGA
AGCCTTCGGCT

As You see always there is a NODE phrase. For each file I would like to change "NODE" for the file indicator from the sample's name: p1/ p2/ p100...pn

The output should look like this for file p1_500sc.fasta:

>p1_60_length_504_cov_1.580902
TATATCGCCGTAGATAGACGAATACGGGATTTTGAAATGACTGATATATTCAAGCACATC
CTGTTCAGTCCAAAATGACATCGGCTTCGATTTCGGGATTGCCCCCGCAAAGTTGTTACA
>p1_61_length_503_cov_4.457447
GATATGGGCCGGTGAGCATGCTGTCACATTTTGGGCAGGTGCCGAGGATTATGAGCTCGT
CTTCCGGCGTCAACGCTTGTTCGTTGCGTCTGGTGATGTGTTCCAGAGCGGCATAGTCGT
>p1_62_length_500_cov_4.037534
CTCGTTCTTGCACTCCTGAATGAAGCGAATCTCTGACGAGGGTATTCCATAATCCTCTAT
CAGCTTACGCTTGATTTCTG
>p1_63_length_500_cov_1.718499
GGAGGATATCAATGCCCTTATACTGCTTCTCAACAGTAGGAGTAATGCGGATACCCAAAA
CGGTCTTAGGGTTCTCCAGCAACTTCATGGCATTCCAACGCTTCAAGTCATCCATGCGGA
AGCCTTCGGCT

For p123_500sc.fasta file

>p123_2_length_456_cov_4.453333
GATATGGGCCGGTGAGCATGCTGTCACATTTTGGGCAGGTGCCGAGGATTATGAGCTCGT
CTTCCGGCGTCAACGCTTGTTCGTTGCGTCTGGTGATGTGTTCCAGAGCGGCATAGTCGT
>p123_356_length_3_cov_4.037908
CTCGTTCTTGCACTCCTGAATGAAGCGAATCTCTGACGAGGGTATTCCATAATCCTCTAT
CAGCTTACGCTTGATTTCTG

The same for any other sample. Sample start from p1, and they can reach p150.

Answer 1

bash + sed solution:

for f in p[0-9]*.fasta; do
    pfx="${f%%_*}"
    sed -i "s/^>NODE/>$pfx/" "$f"
done

---

• pfx="${f%%_*}" - truncate the right part of the filename till the last _ ; assign the result to variable pfx
• sed -i "s/^>NODE/>$pfx/" - substitute >NODE substring with $pfx variable in-place

Answer 2

With GNU awk, you can do it in a single command invocation:

gawk -i inplace -F_ -v OFS=_ '
  BEGINFILE {
    node = FILENAME
    sub(".*/", "", node)
    sub("_.*", "", node)
  }
  $1 == ">NODE" {$1 = ">" node}
  {print}' ./p*_*.fasta

Answer 3

This will loop through each file, assign the everything before the underscore in the filename to the variable sample, then subsitute >NODE for $sample in the file using sed:

for file in /path/to/files/*;do
  sample=$(echo "${file}" | sed s/_.*//)
  sed -i "s/^>NODE/${sample}/" ${file}
done