Extract chromosome 1 - 22 fasta file

Question

[shubhamkumbhar420]

Hello guys I have a fasta file called hg19.fa.gz and it contains chromosomes like

zcat hg19.fa.gz | grep ">"

chr1

chr2

chr3

chrUn_gl000211

chrUn_gl000221

chrUn_gl000214

chrUn_gl000228

chrUn_gl000227

chr1_gl000191_random

chr19_gl000208

and many more

Now I just want chromosomes from 1 to 22 including X and Y

Thank you

Answer 1

[shenwei356]

seqkit grep -i -r -p '^chr[\dXY]+$' h19.fa.gz -o result.fa.gz

[shubhamkumbhar420]

Thank you for your response! It worked !!!

[arsala521]

@shenwei356 This command line is very useful. I wanted to get only canonical chromosomes from female gorilla genome fasta file, i.e. I just wanted chr1, chr2A, chr2B, chr3 - - - chr22, chrX. I edited this command as: seqkit grep -i -r -p '^chr[\dX'2A''2B']+$' gorGor6.fa > output.fa and it worked. I am trying to understand how this is working. I looked into the seqkit grep options -i, -r, and -p. If you can please tell how that string part is exactly working i.e. '^chr[\dXY]+$'. What +$ indicates? Thank you

[shenwei356]

It's a regular expression. Learn more:

• https://geekladysite.wordpress.com/2019/12/12/learn-regex-in-30-minutes/
• https://ryanstutorials.net/regular-expressions-tutorial/regular-expressions-basics.php

And it should be '^chr\w+$' for your case.

[arsala521]

@shenwei356 Hi: Thanks for the help This expression '^chr\w+$' is actually not working. I have fasta file for female gorilla genome having several non-canonical entries. I only want the regular chromosomes i.e. chr1, chr2A, chr2B, chr3, chr4 ..... chr 22, chrX. What expression should I use? TIA

Answer 2

[lmlukoseviciute]

Hello, you I am not really sure if you want to see only chromosome names or the sequences it self. But as of the example you can try.

zcat hg19.fa.gz | egrep "> chr([1-9]$|[1-9]_|1[0-9]|2[0-2]|X|Y)"