Extract sequences based on keywords in headers (no perl or python code)

Question

[Crystal]

Hi All,

I do find some similar questions, but I don't know how to apply the answers to solve my problem.

I guess it's very simple to everyone here.

I have a fast file containing 2000+ sequences with headers like:

>VFG0001 fhaB filamentous hemagglutinin/adhesin [FHA] [Bordetella pertussis]

the content in the last [ ] is the bacteria name.

Also for some other sequences, the header is like

>VFG2304 misL putative autotransporter [MisL] [Salmonella enterica (serovar typhimurium)]

So I just want to extract sequences that have headers containing key word [Escherichia coli].

I also want to extract sequences that have headers containing key word [Salmonella enterica] (I don't care about server within ()) into a separate fasta file.

How can I simply do this on Mac server (maybe not use python or perl, just simple script)?

Thank you

Answer 1

[harold.smith.tarheel]

In Terminal:

grep -A 1 '[Escherichia coli' NAME_OF_FILE > E_coli.fasta

Assumes FASTA; change number to 3 for FASTQ. Will also work in cases (such as the Salmonella example) where strain information is present after the species name.

[Crystal]

Thank you so much. It is very helpful. Just a little modification:

grep -A 1 'Escherichia coli' NAME_OF_FILE > E_coli.fasta

because with [ the terminal said unmatched [.

[Crystal]

Hi ,actually there is a problem with the code. After extract the sequences, only the first line of sequences with header is shown in new file, so all the extracted sequences are not complete.

Well, I figured out sequences of the the original database are multiple lines instead of single line, so I just convert all sequences into a single line file and the problem solved.

[Siva]

Or you could use Awk without needing to convert the multiple lines to single lines

https://infoplatter.wordpress.com/2013/10/15/extracting-specific-fasta-records-from-a-multi-fasta-file/comment-page-1/

[harold.smith.tarheel]

Sorry, I've been spending too much time analyzing short-read sequences, and overlooked multi-line FASTA. Glad you worked out a solution.