Recommended Method for Including Sanger Reads

4.13.4

The best method for including Sanger reads into an assembly is to first run phred with vector screening (or an equivalent basecaller that generates a FASTA file with vector sequence marked as ‘X’ or ‘N’, plus a quality score file). The “.fasta.screen” and “fasta.screen.qual” files produced by phred contain the screening and trimming information needed by the GS De Novo Assembler or the GS Reference Mapper (and it is setup to properly process those reads and the marking and quality information).

There are two ways to include the pairing information for the reads, each of which requires some Perl scripting, because the Paired End information is typically encoded in the accession numbers of the reads. The simplest way is to write a Perl script like the one below. This script is designed to:

•	handle a paired read naming convention where there is only one Paired End library;

•	where the forward and reverse reads have suffixes ".f1" or ".r1"; and

•	where the Paired End reads can be distinguished from follow-up finishing reads by having only alphanumeric characters before the ".f1" or ".r1"):

#!/usr/bin/perl

if ($ARGV[0] =~ /^(\w*)\.([rf])1$/) {

   print "template=$1 dir=$2 library=pairlib\n";

where "$ARGV[0]" will be the accession number for the read, when used in the procedure below.

The regular expression in this script states “^(\w*)\.([rf])1$” matches a word of any length, “\w*”, followed by a period, “\.”, followed by either an ‘r’ or an ‘f’, followed by a ‘1’. The parentheses are there to then extract the characters that matched the word and matched the ‘r’ or ‘f’ and place them into the “$1” and “$2” parameters (so that they can be used in the print statement on the next line). For a local naming convention, customize the Perl regular expression to extract out the template, library and forward/reverse direction strings from the accession, and output them appropriately (see Section 4.13.3) for a description of this output.

If this script is named "accnoPair", then the following procedure can be used to incorporate Sanger reads:

1.	Run phred with vector screening to produce the "fasta.screen" file (plus fasta.screen.qual).

2.	Run the command "fnafile -o sanger.fna -ac accnoPair fasta.screen" to attach the Paired End information to each read, where fnafile calls the accnoPair program for each read, then adds the output of that command to the read's description line.

3.	Give sanger.fna to the assembler or mapper, where it will read and use the pairing information.

The one drawback to this method is that the execution of the Perl script for each read causes the fnafile program to take a couple of minutes to convert 15,000-20,000 reads. An advanced script and slightly different procedure that runs the attachments faster can be used:

If the following Perl script is generated (called "detPairs"):

#!/usr/bin/perl

die "Usage: detPairs screenfile\n" unless defined($ARGV[0]); open(FILE, 
$ARGV[0]) or die "Error:  Unable to open file: $ARGV[0]\n";

while (<FILE>) {

    if (/^\>(\S*)/) {

        my $accno = $1;

        if ($accno =~ /^(\w*)\.([rf])1$/) {

            print "$accno template=$1 dir=$2 library=pairlib\n";

then the following procedure can be used to incorporate Sanger reads:

1.	Run phred with vector screening to get the fasta.screen file.

2.	Run the command "detPairs fasta.screen > sanger.acc" to get a file with the pairing information.

3.	Run the command "fnafile -o sanger.fna -af sanger.acc fasta.screen".

4.	Give the sanger.fna file to the assembler or mapper, where it will read and use the pairing information.

For both of these scripts, see Section 3.4 for a description of the command line structure and arguments of fnafile.