|
31
|
1 #!/usr/bin/perl -w
|
|
|
2 #Filename:
|
|
|
3 #Author: Tian Dongmei
|
|
|
4 #Email: tiandm@big.ac.cn
|
|
|
5 #Date: 2013/7/19
|
|
|
6 #Modified:
|
|
|
7 #Description:
|
|
|
8 my $version=1.00;
|
|
|
9
|
|
|
10 use strict;
|
|
|
11 use Getopt::Long;
|
|
|
12
|
|
|
13 my %opts;
|
|
47
|
14 GetOptions(\%opts,"i=s","g=s","index:s","v:i","p:i","r:s","o=s","h");
|
|
31
|
15 if (!(defined $opts{i} and defined $opts{o} ) || defined $opts{h}) { #necessary arguments
|
|
|
16 &usage;
|
|
|
17 }
|
|
|
18
|
|
|
19 my $filein=$opts{'i'};
|
|
|
20 my $fileout=$opts{'o'};
|
|
|
21 unless ($fileout=~/\/$/) {$fileout.="/";}
|
|
|
22 my $genome=$opts{'g'};
|
|
|
23 my $mis=defined $opts{'v'}? $opts{'v'} : 0;
|
|
|
24 my $hits=defined $opts{'r'}? $opts{'r'} : 25;
|
|
|
25 my $index=defined $opts{'index'} ? $opts{'index'} : "";
|
|
|
26 my $threads=defined $opts{'p'} ? $opts{'p'} : 1;
|
|
|
27
|
|
|
28
|
|
|
29 my $time=&Time();
|
|
47
|
30 my $mapdir=$fileout."/genome_match";
|
|
31
|
31 if(not -d $mapdir){
|
|
|
32 mkdir $mapdir;
|
|
|
33 }
|
|
|
34 chdir $mapdir;
|
|
|
35 ###check genome index
|
|
|
36 if (-s $index.".1.ebwt") {
|
|
|
37 }else{
|
|
|
38 `bowtie-build $genome $genome`;
|
|
|
39 $index="$genome";
|
|
|
40 }
|
|
|
41
|
|
|
42 ### genome mapping
|
|
44
|
43 `bowtie -v $mis -f -p $threads -m $hits -a --best --strata $index $filein --al genome_mapped.fa --un genome_not_mapped.fa --max genome_mapped_Mlimit.fa > genome_mapped.bwt 2> run.log`;
|
|
31
|
44
|
|
|
45 #`convert_bowtie_to_blast.pl genome_mapped.bwt genome_mapped.fa $genome > genome_mapped.bst`;
|
|
|
46
|
|
|
47 sub Time{
|
|
|
48 my $time=time();
|
|
|
49 my ($sec,$min,$hour,$day,$month,$year) = (localtime($time))[0,1,2,3,4,5,6];
|
|
|
50 $month++;
|
|
|
51 $year+=1900;
|
|
|
52 if (length($sec) == 1) {$sec = "0"."$sec";}
|
|
|
53 if (length($min) == 1) {$min = "0"."$min";}
|
|
|
54 if (length($hour) == 1) {$hour = "0"."$hour";}
|
|
|
55 if (length($day) == 1) {$day = "0"."$day";}
|
|
|
56 if (length($month) == 1) {$month = "0"."$month";}
|
|
|
57 #print "$year-$month-$day $hour:$min:$sec\n";
|
|
|
58 return("$year-$month-$day-$hour-$min-$sec");
|
|
|
59 }
|
|
|
60
|
|
|
61 sub usage{
|
|
|
62 print <<"USAGE";
|
|
|
63 Version $version
|
|
|
64 Usage:
|
|
|
65 $0 -i -o
|
|
|
66 options:
|
|
|
67 -i input file# input reads fasta/fastq file
|
|
|
68 -g input file# genome file
|
|
|
69 -index file-prefix #(must be indexed by bowtie-build) The parameter
|
|
|
70 string must be the prefix of the bowtie index. For instance, if
|
|
|
71 the first indexed file is called 'h_sapiens_37_asm.1.ebwt' then
|
|
|
72 the prefix is 'h_sapiens_37_asm'.##can be null
|
|
|
73 -v <int> report end-to-end hits w/ <=v mismatches; ignore qualities,default 0;
|
|
|
74
|
|
|
75 -p/--threads <int> number of alignment threads to launch (default: 1)
|
|
|
76
|
|
|
77 -r int a read is allowed to map up to this number of positions in the genome
|
|
|
78 default is 25
|
|
|
79
|
|
|
80 -o output directory
|
|
|
81
|
|
|
82 -h help
|
|
|
83 USAGE
|
|
|
84 exit(1);
|
|
|
85 }
|
|
|
86
|
