Extract fasta sequences from a file using a list in another file.

Question

[EBP91]

I have a question concerning the extraction of sequences from a fasta file (>7000 sequences) using a reference .txt file with sequence headers. I have been playing around and been looking all over the internet to find a solution for this problem, but surprisingly, nothing really matches what I want to do. So, I have two files:

1) a fasta file which looks like this:

>Zotu1

ACTGACAAAGCA

TGCACGTCATTTT

>Zotu2

ATGCATCAGCATA

TGACCCCCGTTTA

>Zotu10

CGTCGAAAAATTT

CGATACACCCTAT

>Zotu22

CGTACGTCCCCTT

CGATATAATATATA

2) a .txt file with a list of sequence names:

Zotu1

Zotu2

Now, I want to use the .txt file to select sequences from the .fasta file. I have two semi-solutions that do part of the job.

OPTION 1:

cat list.txt | awk '{gsub("_","\\_",$0);$0="(?s)^>"$0".*?(?=\\n(\\z|>))"}1' | pcregrep -oM -f - sequences.fas > newfile.fas

Problem: this function gives me the full sequences, but extracts too many sequences since everything that partially matches the strings in the .txt file will be selected. In this case, it means that also Zotu10 and Zotu22 are selected.

OPTION 2:

grep -A 1 -wFf list.txt sequences.fas > newfile2.fas

Problem: this function correctly selects only the sequences that completely match the strings in the .txt file, but does not return the full fasta sequences, but only the part of the sequence on the first line. An output thus looks like this:

>Zotu1

ACTGACAAAGCA

>Zotu2

ATGCATCAGCATA

I tried combining both solutions but that somehow did not end well. I would be much helped by an elegant solution for this problem, preferably using the codes I already obtained.

Many thanks!

[WouterDeCoster]

Please format your fasta sequences appropriately, using the formatting button. You can do this by selecting the text and clicking the 101010 button. When you compose or edit a post that button is in your toolbar, see image below:

[lakhujanivijay]

I just did that for the OP. But yes from next time he or she should do it. Thanks WouterDeCoster

[EBP91]

Thanks! I adjusted it a bit. It should be fine now.

[venu]

Did you try to google it/ tried any solution?

I copy pasted your title and the first link I got is this: Extract fasta sequences from a large file using a list of names

[EBP91]

Yes, I have been entirely through that thread (and several others) before posting here.

Answer 1

[cpad0112]

with seqtk:

 $ seqtk subseq test.fa test.txt 
>Zotu1
ACTGACAAAGCATGCACGTCATTTT
>Zotu2
ATGCATCAGCATATGACCCCCGTTTA

with grep:

$ grep -w -A 2 -f  test.txt test.fa --no-group-separator
>Zotu1
ACTGACAAAGCA
TGCACGTCATTTT
>Zotu2
ATGCATCAGCATA
TGACCCCCGTTTA

inputs:

$ cat test.fa 
>Zotu1
ACTGACAAAGCA
TGCACGTCATTTT
>Zotu2
ATGCATCAGCATA
TGACCCCCGTTTA
>Zotu10
CGTCGAAAAATTT
CGATACACCCTAT
>Zotu22
CGTACGTCCCCTT
CGATATAATATATA

$ cat test.txt
Zotu1
Zotu2

[alowi33]

grep -w -A 2 -f test.txt test.fa --no-group-separator doesn't work if there are special characters in the header, which is common. Use grep -w -A 2 -Ff test.txt test.fa --no-group-separator instead. -F searchers for a fixed string.

For anyone that only has one sequence per header in their fasta file, use grep -w -A 1 -f test.txt test.fa --no-group-separator instead.

[Arindam Ghosh]

What if the number of sequence lines under each header is different? I don't think so this would work.

[EBP91]

Did the job! Thanks!

Answer 2

[GenoMax]

Step 1: Get faSomeRecords utility from Jim Kent at UCSC. http://hgdownload.soe.ucsc.edu/admin/exe/linux.x86_64/faSomeRecords (Linux link, OS X or source available).

Step 2: Make the file executable

$ chmod u+x faSomeRecords

Step 3: Run faSomeRecords

$ ./faSomeRecords
faSomeRecords - Extract multiple fa records
usage:
   faSomeRecords in.fa listFile out.fa

in.fa = Your sequence file

listfile = file with fasta headers/names (one per line)

out.fa = file to store the result

[fibar]

How can you make it work on piped files from zcat? eg. "zcat sequences.fas.gz | faSomeRecords ..."

[malte.ruehlemann]

You can use /dev/stdin:

zcat sequences.fas.gz | faSomeRecords /dev/stdin listFile out.fa

Answer 3

[finswimmer]

Strange (or funny) the third time today I recommended seqkit.

$ seqkit grep -n -f list.txt sequences.fas > newfile2.fas

should do the job.

fin swimmer

[fibar]

Does it work on piped files from zcat? eg. "zcat sequences.fas.gz | seqkit ...". Also, what is the difference with seqkit subseq from above?

Answer 4

[Joe]

This has been asked a lot, so an existing solution will almost certainly match what you need to do.

You will usually need to linearise your fasta:

awk '/^>/ {printf("%s%s\t",(N>0?"\n":""),$0);N++;next;} {printf("%s",$0);} END {printf("\n");}' < myseqs.fasta > linear.fasta

then:

while read IDS ; do grep "\b$IDS\b" linear.fasta ; done < listofids.txt

You can use \b in the grep command to mark word boundaries so that Seq1 doesn't match Seq11 for example.

Then you can rewrap your sequences (partially) by using tr '\t' '\n'.

The code I use is below (using biopython though, so it is a more robust method).

Answer 5

[Rashedul Islam]

You can use this command:

while read line; do less fastaFile.fasta | seqkit grep -p $line; done < listFile >>out.fasta