This is a test version of Biostars. For the public version, visit https://www.biostars.org.
How to build a command to filter an interval using -grep on Linux

I have a set of data that looks like this:

NK.Chr1:75500000-95000000:28960-29007   NG-unitig0655   97.872  47  1   0   1   47  121009  120963  2.90e-14    80.6
NK.Chr1:75500000-95000000:28960-29007   NG-1DRT-unitig0549  97.872  47  1   0   1   47  623680  623726  2.90e-14    80.6
NK.Chr1:75500000-95000000:28960-29007   NG-1DRT-unitig0278  97.872  47  1   0   1   47  1224581 1224627 2.90e-14    80.6
NK.Chr1:75500000-95000000:28960-29007   NG-1DRT-Chr4    97.872  47  1   0   1   47  8416368 8416414 2.90e-14    80.6
NK.Chr1:75500000-95000000:28960-29007   NG-1DRT-Chr4    97.872  47  1   0   1   47  20041035    20041081    2.90e-14    80.6
NK.Chr1:75500000-95000000:28960-29007   NG-1DRT-Chr4    97.872  47  1   0   1   47  35175472    35175426    2.90e-14    80.6
NK.Chr1:75500000-95000000:28960-29007   NG-1DRT-Chr4    97.872  47  1   0   1   47  56460095    56460049    2.90e-14    80.6

I need to filter the lines in the range of 0-3900000, considering only the numbers before NG.

grep 'NK.Chr1:75500000-95000000:[0-3900000]' NG.1DRT-blast.out > chr1-blast-NG.txt

I tried this code, but it returned all the lines with NK.Chr1:75500000-95000000:, not considering the range.

Anyone knows how to build a proper code for it?

grep linux
• 3,103 views • link
written by Alex S • Alex S

Here [0-3900000] is treated as a regular expression, which means any character between 0 and 3 (0, 1, 2, 3), or 9.

$ seq 10 | grep '[0-3900000]'
1
2
3
9
10  # with '1' and '0'
• 0 views • link
written by shenwei356 shenwei356

3 answers

I think the following should work:

awk -F '[:-]'  '($1~/NK.Chr1/ && $2=75500000 && $3=95000000 && $4<=3900000 && $5<=3900000)'
• 0 views • link
written by Jeremy • Jeremy

It works perfectly!! Thanks a lot :)

• 0 views • link
written by Alex S • Alex S 
awk -F '[ \t:-]'  '(!($4> 3900000 || $5<0))' < in.blast
• 0 views • link
written by Pierre Lindenbaum Pierre Lindenbaum

Hi Alex, Shenwei, and Pierre,

I couldn't figure this out until I saw Pierre's code, but I think Alex is asking for the following:

awk -F '[:-]'  '($4<3900000 && $5<3900000)'
• 0 views • link
written by Jeremy • Jeremy

I really need the NK.Chr1:75500000-95000000: as part of the code. Both codes are returning other Chr groups.

• 0 views • link
written by Alex S • Alex S

Try csvtk:

  1. Extract coordinates (saved as 13th and 14th column):

     $ cat data.tsv \
     | csvtk mutate -Ht -f 1 -p '(\d+)-\d+$' \
     | csvtk mutate -Ht -f 1 -p '\d+-(\d+)$' \
     > tmp.tsv

 $ csvtk dim -H  tmp.tsv 
 file      num_cols   num_rows
 tmp.tsv         14          7

  2. Filtering based on the range and remove the temporary columns:

     cat tmp.tsv \
     | csvtk filter2 -Ht -f '$13 >= 0 && $14 <= 3900000' \
     | csvtk cut -Ht -f 1-12 \
     > result.tsv

All-in-one:

$ cat data.tsv \
    | csvtk mutate -Ht -f 1 -p '(\d+)-\d+$' \
    | csvtk mutate -Ht -f 1 -p '\d+-(\d+)$' \
    | csvtk filter2 -Ht -f '$13 >= 0 && $14 <= 3900000' \
    | csvtk cut -Ht -f 1-12 \
    > result.tsv
• 0 views • link
written by shenwei356 shenwei356

Log in to answer this question.

More posts like this

No similar posts found.