基因组重测序的unmapped reads assembly探究 【直播】我的基因组86

在前面的直播基因组系列，我们讲解过那些比对不少我们人类的参考基因组序列的数据，其实可以细致的进行探究。直播】我的基因组（十五）:提取未比对的测序数据这里主要参考这篇文章的图4：http://www.nature.com/ng/journal/v42/n11/figtab/ng.691F4.html

组装的contig注释到物种这是2010年发表于nature genetics杂志的Whole-genome sequencing and comprehensive variant analysis of a Japanese individual using massively parallel sequencing 虽然文章选择的是SOAPdenovo,ABySS,Velvet这3款软件来进行组装，但毕竟是2010年的文章了，现在其实有更好的选择，比如Minia选择Minia工具来组装Minia软件也是基于de Bruijn图原理的短序列组装工具，优于以前的ABySS和SOAPdenovo，关键是速度非常快，十几分钟就OK了，不消耗计算机资源，所以这里就选择它啦。下载安装Minia安装官网的指导说明书下载二进制版本即可，代码如下：## Download and install Minia# http://minia.genouest.org/cd ~/biosoftmkdir Minia &&  cd Miniawget https://github.com/GATB/minia/releases/download/v2.0.7/minia-v2.0.7-bin-Linux.tar.gztar -zxvf minia-v2.0.7-bin-Linux.tar.gz~/biosoft/Minia/minia-v2.0.7-bin-Linux/bin/minia --help## eg: ./minia -in reads.fa -kmer-size 31 -abundance-min 3 -out output_prefix软件使用方法也非常简单，就一行命令，其中最佳 -kmer-size需要用KmerGenie来确定。使用step1:提取比对失败的readssamtools view -f4 jmzeng_recal.bam |perl -alne '{print "\@$F[0]\n$F[9]\n+\n$F[10]" }' >unmapped.fqperl ~/biosoft/PRINSEQ/prinseq-lite-0.20.4/prinseq-lite.pl -verbose -fastq unmapped.fq -graph_data unmapped.gd -out_good null -out_bad nullperl ~/biosoft/PRINSEQ/prinseq-lite-0.20.4/prinseq-graphs.pl -i unmapped.gd -png_all -o unmappedperl ~/biosoft/PRINSEQ/prinseq-lite-0.20.4/prinseq-graphs.pl -i unmapped.gd -html_all -o unmappedcd ~/data/project/myGenome/gatk/jmzeng/unmapped共31481084/4=7870271，仅仅是7.8M的readsInput InformationInput file(s):unmapped.fqInput format(s):FASTQ# Sequences:7,870,271Total bases:1,180,540,650step2: 用KmerGenie确定kmer值KmerGenie estimates the best k-mer length for genome de novo assembly.KmerGenie predictions can be applied to single-k genome assemblers (e.g. Velvet, SOAPdenovo 2, ABySS, Minia).## http://kmergenie.bx.psu.edu/cd ~/biosoftmkdir KmerGenie &&  cd KmerGeniewget http://kmergenie.bx.psu.edu/kmergenie-1.7044.tar.gztar zxvf kmergenie-1.7044.tar.gzcd kmergenie-1.7044makepython setup.py install --user~/.local/bin/kmergenie --helpcd ~/data/project/myGenome/gatk/jmzeng/unmapped~/.local/bin/kmergenie unmapped.fqstep3: 运行Miniacd ~/data/project/myGenome/gatk/jmzeng/unmapped~/biosoft/Minia/minia-v2.0.7-bin-Linux/bin/minia  -in unmapped.fq -kmer-size 31 -abundance-min 3 -out output_prefix7.8M的reads组装之后有272007条contigs组装之后：Prinseq v0.20.4 was used to calculate assembly statistics, including N50 contig size, GC contentcd ~/data/project/myGenome/gatk/jmzeng/unmappedperl ~/biosoft/PRINSEQ/prinseq-lite-0.20.4/prinseq-lite.pl -verbose -fasta output_prefix.contigs.fa  -graph_data contigs.gd -out_good null -out_bad nullperl ~/biosoft/PRINSEQ/prinseq-lite-0.20.4/prinseq-graphs.pl -i contigs.gd -png_all -o contigsperl ~/biosoft/PRINSEQ/prinseq-lite-0.20.4/prinseq-graphs.pl -i contigs.gd -html_all -o contigsperl ~/biosoft/PRINSEQ/prinseq-lite-0.20.4/prinseq-lite.pl -verbose -fasta output_prefix.contigs.fa  -stats_assembly就是给出一些指标，如下；stats_assembly    N50 176stats_assembly    N75 113stats_assembly    N90 78stats_assembly    N95 70Input InformationInput file(s):output_prefix.contigs.faInput format(s):FASTA# Sequences:272,007Total bases:44,868,011Length DistributionMean sequence length:164.95 ± 204.44 bpMinimum length:63 bpMaximum length:10,187 bpLength range:10,125 bpMode length:150 bp with 16,461 sequences然后用RNA-SEQ数据来比对验证！ 以后再讲把组装好的contigs拿去NCBI做blast看看物种分布,Distribution of top nucleotide BLAST hits by species from the NCBI nr database for 1000 random contigs in the assembly！其实上面的prinseq软件也简单的给出了一个污染物种分布情况表，但是这个原理不一样。以后再讲