basicLimma(AgiMicroRna)
basicLimma()所属R语言包：AgiMicroRna
Linear models Using limma
使用limma线性模型
译者：生物统计家园网 机器人LoveR
描述----------Description----------
Differential expression analysis using the linear model features implemented in the limma package. A linear model is fitted to each miRNA gene so that the fold change between different experimental conditions and their standard errors can be estimated. Empirical Bayes methods are applied to obtain moderated statistics
采用线性模型在limma包实现的功能差异表达分析。线性模型被安装到每个miRNA基因，使不同的实验条件和标准误差之间的fold change，可估计。经验Bayes方法适用于获得主持的统计
用法----------Usage----------
basicLimma(eset, design, CM,verbose = FALSE)
参数----------Arguments----------
参数：eset
ExpressionSet containing the processed log-expression values
ExpressionSet包含的处理log表达式的值
参数：design
design matrix
设计矩阵
参数：CM
contrast matrix
对比矩阵
参数：verbose
logical, if TRUE prints out output
逻辑，如果TRUE打印出输出
Details
详情----------Details----------
In our data example (see the target file in Table 1 in vignette), we have used a paired design (by subject) to assess the differential expression between two treatments B and C vs a control treatment A. That is, we want to obtain the microRNAS that are differentially expressed between conditions A vs B and A vs C. The linear model that we are going to fit to every miRNA is defined by equation: y = Treatment + Subject + error term. This model is going to estimate the treatment effect and then, the comparison between the different treatments are done in terms of contrasts between the estimates of the treatment effects. To fit the model, we need first to define a design matrix. The design matrix is an incidence matrix that relates each array/sample/file to its given experimental conditions, in our case, relates each file to one of the three treatments and with its particular subject. If treatment is a factor variable, we can define de desing matrix using model.matrix(~ -1 + treatment + subject). Then the linear model can be fitted using fit=lmFit(eset,design). This will get the treatment estimates for each microRNA in the eset object:
在我们的数据的例子（见目标文件中的表1中的小插曲），我们已经使用配对设计（主题），两种治疗B和C对控制治疗答：这是评估之间的差异表达，我们要获得条件A对B，A与C的线性模型，我们要以适应每个miRNA的方程定义之间的差异表达的microRNA：Y = +主题+误差项处理。该模型估计的治疗效果，然后，不同的处理方法之间的比较，在治疗效果的估计之间的对比。以适应模型，我们首先需要定义一个设计矩阵。设计矩阵是一种发病率矩阵，关乎每个阵列/样本/文件，其给定的实验条件，在我们的例子中，涉及每个文件的三种治疗方法之一，其特定主题。如果治疗是一个因素变量，我们可以定义去德兴矩阵使用model.matrix（~（-1）+治疗+主题）。然后可以安装使用适合lmFit（ESET，设计）的线性模型。这将让每个在ESET对象的microRNA治疗估计：
treatmentA treatmentB treatmentC subject2 hsa-miR-152 7.5721 7.656 7.566 -0.1157 hsa-miR-15a* 0.9265 1.066 1.211 -0.2242 hsa-miR-337-5p 6.2448 7.298 7.084 -0.4489
treatmentA treatmentB treatmentC subject2 HSA-MIR-152 7.5721 7.656 7.566 -0.1157 * 0.9265 1.066 1.211 -0.2242 HSA-的miR-15A HSA-MIR-337-5P 6.2448 7.298 7.084 -0.4489
We can define the contrasts of interest using a contrast matrix as in CM=cbind(MSC\_AvsMSC\_B=c(1,-1,0), MSC\_AvsMSC\_C=c(1,0,-1))
我们可以定义的反差对比矩阵的兴趣在CM = cbind（海安\ _AvsMSC \ _B = C（1，-1,0），海安\ _AvsMSC \度C = C（1,0，-1））
And then, we can estimate those contrats using fit2=contrasts.fit(fit,CM). Finally, we can obtain moderated statistics using fit2=eBayes(fit2).
然后，我们可以估算那些contrats使用FIT2 = contrasts.fit（适合CM）。最后，我们可以得到使用FIT2 = eBayes（FIT2）主持的统计数据。
The function 'basicLimma' implemented in AgiMicroRna produces the last fit2 object, that has in fit2\$coeff the M values, in fit\$t the moderated-t statistic of the contrasts, and in fit2\$p.value the corresponding p value of each particular contrasts. Be aware that these p values must be corrected by multiple testing.
函数basicLimma在AgiMicroRna实施产生最后FIT2的对象，有FIT2 \ $ coeff M值，在合适的\ $ T-T放缓的对比统计，在FIT2 \ $ p.value相应的P值的每一个具体的对比。要知道，这些p值必须纠正多个测试。
MSC\_AvsMSC\_B MSC\_AvsMSC\_C hsa-miR-152 0.67567761 0.977326746 hsa-miR-15a* 0.68019442 0.413657270 hsa-miR-337-5p 0.03737814 0.075248741
海安\ _AvsMSC \ _B海安\ _AvsMSC \度C HSA-MIR-152 0.67567761 0.977326746 HSA的miR-15A * 0.68019442 0.413657270 HSA-的miR-337-5P 0.03737814 0.075248741
See limmaUsersGuide() for a complete description of the limma package.
看到一个完整的描述的limma包limmaUsersGuide（）。
值----------Value----------
An MArrayLM object of the package limma
一个包limma MArrayLM对象
作者（S）----------Author(s)----------
Pedro Lopez-Romero
参考文献----------References----------
'Bioinformatics and Computational Biology Solutions using R and Bioconductor'. R. Gentleman, V. Carey, S. Dudoit, R. Irizarry, W. Huber (eds), Springer, New York, pages 397–420.
diferential expression in microarray experiments. Statistical Applications in Genetics and Molecular Biology, Vol. 3, No. 1, Article 3. http://www.bepress.com/sagmb/vol3/iss1/art3
参见----------See Also----------
An 'RGList' example containing proccesed data is in ddPROC and an overview of how the processed data is produced is given in filterMicroRna. The ExpressionSet object can be generated using esetMicroRna
一个“RGList”含有proccesed数据的例子是在ddPROC和处理的数据是如何产生的概述filterMicroRna。可以生成ExpressionSet对象，使用esetMicroRna
举例----------Examples----------
## Not run: [＃无法运行：]
data(targets.micro)
data(ddPROC)
esetPROC=esetMicroRna(ddPROC,targets.micro,makePLOT=FALSE,verbose=FALSE)
levels.treatment=levels(factor(targets.micro$Treatment))
treatment=factor(as.character(targets.micro$Treatment),
levels=levels.treatment)
levels.subject=levels(factor(targets.micro$Subject))
subject=factor(as.character(targets.micro$Subject),
levels=levels.subject)
design=model.matrix(~ -1 + treatment + subject )
CM=cbind(MSC_AvsMSC_B=c(1,-1,0,0),
MSC_AvsMSC_C=c(1,0,-1,0))
fit2=basicLimma(esetPROC,design,CM,verbose=TRUE)
names(fit2)
head(fit2$coeff)
head(fit2$p.value)
plot(fit2\$Amean,fit2$coeff[,1],xlab="A",ylab="M")
abline(h=0)
abline(h=c(-1,1),col="red")
plot(fit2$coeff[,1],fit2$p.value[,1], xlab="M",ylab="p value")
## End(Not run)[＃结束（不运行）]
转载请注明:出自 生物统计家园网(http://www.biostatistic.net)。
注：
注1：为了方便大家学习，本文档为生物统计家园网机器人LoveR翻译而成，仅供个人R语言学习参考使用，生物统计家园保留版权。
注2：由于是机器人自动翻译，难免有不准确之处，使用时仔细对照中、英文内容进行反复理解，可以帮助R语言的学习。
注3：如遇到不准确之处，请在本贴的后面进行回帖，我们会逐渐进行修订。
|
发表于 2012-2-25 11:37:21
