Check If A File Is In Fasta Format

Question

[sana.atique.khan]

I am trying to write a code which asks for a file (if the first time, an invalid filename is given, it asks for file 5 times until exhausting), then it checks if the file is in fasta format.

how to code that? I have the following code so far.

#!/usr/bin/perl -w
#A program that asks for a file, opens it if file exists and check
#if the file is in FASTA format
use strict;
#get data from a file
my @file = openfile();

#open file
#subroutines
sub openfile {
my $filename;
my $x;
    my $datafile;
    my $file;

for  ($x = 0; $x<5; $x++) {
print "\n\nPlease enter file name: ";
chomp ($filename = <STDIN>);

if (-e $filename) {
print "File found!\n\n";
    exit;
         } else {
        if ($x<4) {
        print "Invalid file name!\n\n";
        } else { 
                print "Five tries were unsuccessful! Please check and try again!\n\n";
                        }
                    }
                }
        return;
        }

Answer 1

[Neilfws]

First, there is no need to reinvent the wheel. Use the SeqIO module from Bioperl:

#!/usr/bin/perl -w
use strict;
use Bio::SeqIO;

my $seqio = Bio::SeqIO->new(-file => "myfile.fa", -format => "fasta");
while(my $seq = $seqio->next_seq) {
  # do stuff with sequences...
}

If the fasta file is invalid, this code will throw an exception, for example:

------------- EXCEPTION: Bio::Root::Exception -------------
MSG: The sequence does not appear to be FASTA format (lacks a descriptor line '>')

Second, don't waste time checking for multiple incorrect attempts. Once is enough :)

[sana.atique.khan]

Thanks!!! I will try it out!

Answer 2

[Pierre Lindenbaum]

You could create a simple grammar for FASTA using GNU-Bison:

%{
#include <stdio.h>
#include <stdlib.h>
#include <ctype.h>

int yylex();
int yyerror( char* message);
%}
%error-verbose
%token LT OTHER SYMBOL CR
%start input
%%

input:   input  sequence | optspaces sequence;
sequence: head body optspaces;
head: LT anylist CR | LT CR;
anylist: anylist any | any;
any: LT | OTHER | SYMBOL;
body: symbols CR | body symbols CR ;
symbols: symbols symbol | symbol ;
symbol: SYMBOL;
optspaces: | crlist;
crlist: crlist CR | CR;

%%
int yyerror( char* message)
    {
    fprintf(stderr,"NOT A FASTA %s\n",message);
    exit(EXIT_FAILURE);
    return -1;
    }
int yylex()
    {
    int c=fgetc(stdin);
    switch(c)
        {
        case EOF: return c;
        case '>' : return LT;
        case '\n' : return CR;
        default: return isalpha(c)?SYMBOL:OTHER;
        }
    }

int main(int argc, char** argv)
    {
    return yyparse();
    }

and use it to test if a file is a fasta file:

#compile
bison fasta.y
gcc -Wall -O3 fasta.tab.c

#test
$ ./a.out < ~/file.xml
NOT A FASTA syntax error, unexpected OTHER, expecting LT

$ ./a.out < ~/rotavirus.fasta
$

[Jeremy Leipzig]

coming up with a set of these for popular formats would actually be a nice addition to any Makefile or similar pipeline: "BioValidators by Pierre"

Answer 3

[ngsgene]

Unless I am reading this too straightforwardly, you simply need to add an if to test if a file is in fasta format with condition if($l ~=/>/)

if the line ($l) contains > you're good to go.

while(my $l = <DAT>) {
    chomp $l;
    if ($l ~= />/) {
    do this
    }
    elsif ($ !~ />/) {
    do this
    }
}

[Neilfws]

This is a little simplistic. You should at least check whether > is at the start of the first line, using /^>/. Also, there should be a check for no space after >. And then there is the problem of valid sequence lines.

[ngsgene]

True, its more useful for parsing a fasta file when you know its a fasta file.