Removing ">" before coordinates and separate the coordinates in a single string format into a bed file

Question

[Kai_Qi]

I have a fasta file that contains the coordinates and the sequence of the coordinates:

head my.fasta
>16:23107820-23108019(+)
GTACGGCGCTCCCGGGGCGGCCGGTGGCCTGTAGTCAAGGTCACTAGGACCCGCGTTGAGGTGGGTTGCTTGGCGGCCACACTGCAGGTATGCGGGCTTTTTCTTAGGGCACACACTTCTCCTTGTGCCCTTCGAGAAGCTTCCATGATGGTAAGACTCCAGATGTTGGGGAGACAGGACGGATACAAGAACGGAGTAT
>14:54909471-54909670(-)
GTAAGTGGCACCCTGCCAGAGATCCCTCTCTGCCCTGGGTCTCATGCCTTCCTTTCTGCACCTCCAGACAATTTCTGCTGCCCCTAGGTCCCAGATTTCAGCTGTCCAGATGTCCAGGCCTTTTAAAGGGTCTAGGCAGGGGGTCCTACTGCTCACACAGTCCTCCCACTGGCTGTTATGTTTAAAATCCTAACCTGGC
>7:127020805-127021004(-)
GTAGGTGTGGACGACAGACAGCTGGGTGGCATGAGAATGCAGGTGCCAGGCGAACTAGAGGGTGGTGCTGGGTGCGTCGTACCATCGGGAGAAGATCCCCTCCCCCTCAGCCTCTGCTGAAAGCAACAAGGGAACCCCTAAAAGAAGGGCTAAGAAGGTATGCACAAGATACTGGGTCTTCCCCAAGAATGGGGCTGGA
>X:20848619-20848818(+)
GTGAGGGCAGGCCCGGTAGGGTTCGGGTTTTGGAGCGGCTGCGGGACCCGGGTATGAAGTCCAGACCGAAAGCTCAGCTCCAAGATGCTTCCGTCTGAATCTCAGCGTTCTCCCGCCCGGAACCAAAGGAGTGGTTTGACCAGGGCGAGACCGTCGTCATCGACCGTGGGAGTGGATGGAGGAGTCGGCCTGCAGGCTG
>1:75547398-75547597(+)
GTGGGTAGCCTGGGGACCCCTAGCACCCCAGCCTTCACCACCATCACCTTCATCGCCACCATTACTGCGCTCACCTCCGGCTTGATCACTCAGTGTCATCCTGTGCTGGACGCTGTGCTGGGCCACCATGCCATGTTAAGTCATCCTGCCTCTCATACCATCATCACCTTGTTCACCTGTCAGGGGAGATGTAGGGGAG

I used grep "^>" my.fasta > mycoord.csv and grep "^>" my.fasta > mycoord.bed to extract the coordinates. Now I have seen them there:

$head mycoord.csv (or head mycoord.bed)

>16:23107820-23108019(+)
>14:54909471-54909670(-)
>7:127020805-127021004(-)
>X:20848619-20848818(+)
>1:75547398-75547597(+)
>11:102777648-102777847(+)
>7:25314905-25315104(+)
>2:180025312-180025511(+)
>7:30533903-30534102(-)
>X:8128769-8128968(-)

My question is how to remove the ">" before each coordinates and how can I make the coordinates into several columns so that I can get the gene name using the coordinates and strand information (I don't know how to express the ">" stuff so that I when I searched how to remove ">" in coordinates I almost got nothing)?

Thanks,

[Kai_Qi]

I found the answer to the first part: I used $ sed 's/>//' mycoord.csv > mycoord_1.csv to remove ">"

$ head mycoord_1.csv
16:23107820-23108019(+)
14:54909471-54909670(-)
7:127020805-127021004(-)
X:20848619-20848818(+)
1:75547398-75547597(+)
11:102777648-102777847(+)
7:25314905-25315104(+)
2:180025312-180025511(+)
7:30533903-30534102(-)
X:8128769-8128968(-)

Answer 1

[cpad0112]

several ways: with grep:

$ grep -oP '(?<=\>).*' test.txt

16:23107820-23108019(+)
14:54909471-54909670(-)
7:127020805-127021004(-)
X:20848619-20848818(+)
1:75547398-75547597(+)

with seqkit (https://github.com/shenwei356/seqkit):

$ seqkit seq -n test.txt

16:23107820-23108019(+)
14:54909471-54909670(-)
7:127020805-127021004(-)
X:20848619-20848818(+)
1:75547398-75547597(+)

To convert to space separated coordinates:

$ seqkit replace -ip '(\w+):(\w*)-(\w*).([+,-]).' -r '$1 $2 $3 $4' test.txt | seqkit seq -n

16 23107820 23108019 +
14 54909471 54909670 -
7 127020805 127021004 -
X 20848619 20848818 +
1 75547398 75547597 +

To separate tab separated coordinates:

$ sed -n '/^>/ s/^>\(\w\+\)\W\(\w\+\)\W\(\w\+\).\([+,-]\)./\1\t\2\t\3\t\4/p' test.txt

16      23107820        23108019        +
14      54909471        54909670        -
7       127020805       127021004       -
X       20848619        20848818        +
1       75547398        75547597        +

Output is in 1 indexed, non-bed format. Convert to bed format (0 based start) and intersect with appropriate gtf.

[Kai_Qi]

The solutions worked. Thanks a lot.