——标题:“5。“作者:”Martin Morgan (martin.morgan@roswellpark.org)
罗斯威尔帕克癌症研究所，布法罗，纽约
5 - 9 October, 2015" output: BiocStyle::html_document: toc: true toc_depth: 2 vignette: > % \VignetteIndexEntry{5。} % \VignetteEngine{knitr::rmarkdown}——' ' ' {r style, echo = FALSE, results = 'asis'} BiocStyle::markdown() options(width=100, max.print=1000) knitr::opts_chunk$set(eval=as.logical(Sys. logical(Sys. php)) $set(eval=as.logical(Sys. php)采用“KNITR_EVAL”,“真正的”)),缓存= as.logical (Sys。采用“KNITR_CACHE”,“真正的”)))``` ```{r setup, echo=FALSE, messages=FALSE, warnings=FALSE} suppressPackageStartupMessages({library(GenomicFiles) library(BiocParallel)})本课程的材料需要Rversion 3.2和Bioconductor version 3.2 {R configuration -test} stopifnot(getRversion() >= '3.2' && getRversion() < '3.3'， BiocInstaller::biocVersion() == "3.2")#限制-只输入必要的数据，例如:' ScanBamParam() ' - ' which ':感兴趣的基因组范围- ' what ':##迭代——读取整个文件，但以块为单位——块大小足够小，可以轻松地放入内存，块大小足够大，可以从_R_的向量化操作中受益——一次10k到1M记录——例如，' BamFile(…， yieldSize=100000)的迭代编程模型- yield_数据块-将输入数据映射为方便的表示，通常将输入汇总为简化的形式-例如，对齐读坐标计数感兴趣的重叠区域-例如，对齐读序列到GC内容- _reduce_跨映射块-使用' GenomicFiles::reduceByYield()` ```{r iteration} library(GenomicFiles) yield <- function(bfl) {## input a chunk of alignments library(GenomicAlignments) readGAlignments(bfl，} map <- function(aln) {## Count G or C nucleotides per read library(Biostrings) gc <- letterFrequency(mcls (aln)$seq， " gc ") ## summary read number with 0, 1，…G或C核苷酸表(1 + gc, 73) # max。read length: 72} reduce <- ' +` ``` - Example ' ' {r iterator -doit}库(RNAseqData.HNRNPC.bam.chr14) fls <- RNAseqData.HNRNPC.bam. bam. xml)chr14_BAMFILES bf <- BamFile(fls[1], yieldSize=100000) gc <- reduceByYield(bf, yield, map, reduce) plot(gc, type="h", xlab="GC Content per Aligned Read", ylab="Number of Reads") ``` ## Parallel evaluation - Cores, computers, clusters, clouds - Generally, requires memory management techniques like restriction or iteration -- parallel processes competing for shared memory - Many problems are _embarassingly parallel_ -- `lapply()`-like -- especially in bioinformatics where parallel evaluation is across files - Example: GC content in several BAM files ```{r parallel-doit} library(BiocParallel) gc <- bplapply(BamFileList(fls), reduceByYield, yield, map, reduce) library(ggplot2) df <- stack(as.data.frame(lapply(gc, cumsum))) df$GC <- 0:72 ggplot(df, aes(x=GC, y=values)) + geom_line(aes(colour=ind)) + xlab("Number of GC Nucleotides per Read") + ylab("Number of Reads") ``` # Resources Acknowledgements - Core (Seattle): Sonali Arora, Marc Carlson, Nate Hayden, Jim Hester, Valerie Obenchain, Hervé Pagès, Paul Shannon, Dan Tenenbaum. - The research reported in this presentation was supported by the National Cancer Institute and the National Human Genome Research Institute of the National Institutes of Health under Award numbers U24CA180996 and U41HG004059, and the National Science Foundation under Award number 1247813. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health or the National Science Foundation. ## `sessionInfo()` ```{r sessionInfo} sessionInfo() ```