生物信息 perl 脚本实战
索引
1.统计fasta、fa和fastq文件的长度,统计fastq的reads个数,单个reads长度,reads总长度;统计fasta文件中contig的个数,列出名称,单条的长度,以及总长度。
2.1局部组装:创建目录,将比对好的reads按100k为单位,用samtools切,并用awk工具提起reads,分别存放在对应文件夹内
2.2局部组装:用得到的reads_name,去原始的下机reads里提取reads序列
2.3局部组装:使用canu/soapdenovo按窗口组装
基本套路:
1.使用传参数模块
use Getopt::Long; my ($first, $second, $third, $fourth); GetOptions( "first=s" => \$first, "second=s" => \$second, "third=s" => \$third, "fourth=s" => \$fourth, ) or die $!; print "$first-$second-$third-$fourth\n";
模块名是Getopt::Long,不能写错;使用时,得用GetOptions函数,函数中间是个哈希结构,将--first传给$first变量,以此类推。
这个模块有点复杂,具体参照:在Perl中使用Getopt::Long模块来接收用户命令行参数
2.读表格式文件,切分,取出目的字段,建立数据结构
open IN, "test.txt" or die $!; while(<IN>){ chomp; @list = split / /; next if $list[0] =~ /\bchar1\b|char4|char6/; $name = $list[2]; $value = $list[1]; $hash{$name} = $value; foreach (keys %hash){ print "$_ => $hash{$_}\n"; } } close(IN);
3.将原始的下机reads读到哈希结构里,reads名字作为键,其碱基序列作为哈希的值
#Store the original fasta file into a hash named %hash_reads. my %hash_reads; my $name; open IN,"<",$original_fasta or die $!; while(<IN>){ chomp; if(/^>(.*)$/){ $name=$1; $hash_reads{$name}=""; }else{ $hash_reads{$name} .= $_; } } close IN;
1.统计fasta、fa和fastq文件的长度,统计fastq的reads个数,单个reads长度,reads总长度(主要是统计总长度,其他在Linux下很简单就实现了);统计fasta文件中contig的个数,列出名称,单条的长度,以及总长度。
思路整理:这是个典型的逐行读文件,取字段,计算长度问题。
fastq简单:四行一轮,解决办法很多,可以逐行读取,逐行匹配;也可以一次读两行;输出也少,单reads长度,reads数量,reads长度总和,也没有其他有价值的信息。
fasta略微复杂:没有规律,因为序列被切成短的,只能逐行读,逐行匹配;逐行读有个问题,怎么检测结束?(这是逐行读的一个最大缺陷,你无法操纵最后一次!!!)因此,只能把最后一次放在读循环的外面,每次输出的点就在匹配title那个地方。
代码如下:
#!/usr/bin/perl #Author:zxlee #Function: compute the length of fastq or fasta, fa. #usage: `perl script_name fastq/fasta/fa_file_name`, it will show the total length, also a detail file. use strict; use warnings; my $infile = shift; #give 1st default para to it, you can go on shift to get the 2st para open IN, "<$infile" or die $!; open OUT, ">./result_len.txt" or die $!; our $total_len = 0; our $seq_num = 0; our $len; if($infile =~ /fastq$/){ while(<IN>){ next if not /^@\S+/; my $seq = <IN>; #you cannot use $_ here!!! chomp($seq); $seq_num += 1; $total_len += length($seq); print OUT "\nreads_len = $total_len\n" if $seq_num == 1; } print OUT "Total num of reads is $seq_num\n"; } elsif($infile =~ /(fasta|fa)$/){ # easy way, no "OR" my $chr_len = 0; while(<IN>){ chomp; my $line = $_; if ($line =~ /^>(\S+)/){ print OUT "$chr_len\n" if $chr_len != 0; print OUT "$1\t"; $chr_len = 0; }else{ $len = length($line) if $total_len == 0; $chr_len += length($line); $total_len += length($line); } } print OUT "$chr_len\n"; print OUT "one line has $len\n"; } print "The total length is $total_len\n"; close(IN); close(OUT);
总结:若直接读到数组里,会大大降低操作的复杂度,思维也极容易理解,但是非常耗费计算资源,看数据量决定用哪一种方法。
此脚本主要值得学习的地方:
1.shift的用法,用于逐个取得命令行的参数,让脚本得到封装,不必每次使用时都在里面改文件路径。
2.my、our、local的用法,一般只用my,my管本层及以内的所有层次,my写在最外面就等价于our。
3.$_什么时候会被改变,不是异想天开,while(<IN>)等价于while($_=<IN>),但是单独<IN>就没有这个效果,这个就是几种局限的用法,不要脑洞大开,随意篡改语法。
3.或的模式匹配,/(fasta|fa)$/),$就不能写在里面,不知道为什么,记住就好了。
4.算法流程的问题,写程序之前一定要明晰大致的算法,写的时候要是没有算法框架的支持,那你就痛苦死了。
2.1局部组装,创建目录,将比对好的reads按100k为单位,用samtools切,并用awk工具提起reads,分别存放在对应文件夹内
#!/usr/bin/perl #Author: zxlee #Function: ref, origin reads, bwa_sam_filterd_sored.bam, I want split reads per 100k, for local assembly. #Usage: {perl script_name --size=100000 --bam=your_sorted.bam --jobs=200 --outpath=/split_reads_name --sample=DHX_Y255} use strict; use warnings; use Getopt::long; my ($size, $bam, $outpath, $sample_name, $jobs); # get the perl script para GetOptions( "size=s" => \$size, "bam=s" => \$bam, "jobs=s" => \$jobs, "sample=s" => \$sample_name, "outpath=s" => \$outpath, ) or die $1; # my hash, chr => chr_size my %hash = ( 'Chr1' => 43270923, 'Chr2' => 35937250, 'Chr3' => 36413819, 'Chr4' => 35502694, 'Chr5' => 29958434, 'Chr6' => 31248787, 'Chr7' => 29697621, 'Chr8' => 28443022, 'Chr9' => 23012720, 'Chr10' => 23207287, 'Chr11' => 29021106, 'Chr12' => 27531856, ); # iterate each chr, mkdir, create pbs, submit pbs foreach my $chr (keys %hash){ my $split_bam_dir = $outpath."/${chr}_split_bam/"; # dir storage the split bam, abs dir my $split_pbs_dir = $outpath."/${chr}_split_pbs/"; my $split_log_dir = $outpath."/${chr}_split_log/"; my $split_read_names = $outpath."/${chr}_split_read_names/"; system("mkdir $split_bam_dir") if not -e $split_bam_dir; #create dir system("mkdir $split_pbs_dir") if not -e $split_pbs_dir; system("mkdir $split_log_dir") if not -e $split_log_dir; system("mkdir $split_read_names" if not -e $split_read_names; # create split psb for(my $start=1; $start<$hash{$chr}; $start+=$size){ my $end = $start + $size; my $startHK = int($start/$size); # for name convinient my $endHK = int($end/$size); my $bam_name = "$split_bam_dir/${sample_name}_${chr}_${startHK}hk_${endHK}hk.bam"; my $read_name = "$split_read_names_dir/${sample_name}_${chr}_${startHK}hk_${endHK}hk_reads_name.txt"; my $pbs_name = "$split_pbs_dir/${sample_name}_${chr}_${startHK}hk_${endHK}hk.pbs"; open PBS, ">$pbs_name"; print PBS <<SET; #PBS -N split_${startHK}_${endHK} #PBS -j oe #PBS -l nodes=1:ppn=1 #PBS -q low #PBS -o $split_log_dir hostname date cd \$PBS_O_WORKDIR export LD_LIBRARY_PATH=\$LD_LIBRARY_PATH:/public/software/htslib-1.3/lib /public/software/samtools-1.3/bin/samtools view -b $bam $chr:$start-$end -o $bam_name /public/software/samtools-1.3/bin/samtools view $bam_name | awk '{arr[\$1]} END{for(key in arr) {print key} }' > $read_name date SET close PBS; while(1){ my @run_jobs = `qstat -u zxli | grep "split`; if(@run_jobs < $jobs){ lase; }else{ sleep(1); } } system("qsub $pbs_name"); last; # for test } last; # for test }