Dear Chintanu ad 1.: the assumption is that the arrays that you compare study similar biological system, e.g. the same cell line treated with different drugs, such that the gene expression differences for most genes are relatively minor. The assumption would not hold if you compare, say, a cow liver to human glioblastoma. ad 2.: I am not sure I understand your question, and maybe you need to precisify it. The heatmap is a data exploration tool, there is no single right way to read it. For instance, in Fig 3.2. that you pointed to, I'd say there are three clusters (batches) of arrays, with two of them a bit more similar to each other than to the third. Best wishes Wolfgang On Jul/31/10 8:04 AM, Chintanu wrote: > Hi, > > I am unable to find answers to two queries based on a paper > (*arrayQualityMetrics--a > bioconductor package for quality assessment of microarray data*) and a book > (*Bioconductor Case Studies*). Wonder if anybody could help: > > 1. I understand that a heatmap is a representation of distances between the > arrays, where the median of the absolute values of the difference between > each array-pair is considered as a measure of distance. That is, *Mxi *= *zi > *+ *?**xi** *+ *?**xi* where *Mxi *and *Myi *represents the *M*-value of the > *i*th probe on the *x* and *y* array. For *?xi *, which represents > differential expression effects, it is said that the majority of its values > are negligibly small (or close to zero). Why is it so ? Is this because > median/mean gene may possibly be assumed to be not expressed !! Please let > me know. > > 2. How to interpret a heatmap ? Let's say there is a heatmap of same > framework as that of Fig. 3.2, page 30 of the book, *Bioconductor Case > Studies* (http://tinyurl.com/2ervgpn). I understand that a heatmap provides > the distance matrix entries (dxy), and in expectation, all the values > of dxyare the same, namely 2-times the > *?**xi*. However, what are the key elements in a figure such as Fig. 3.2 ? > If the key elements are the tree-diagram on the top and the colour scale on > the right, then how to interpret them in the sense that (depending on the > diagram,) a set of one/more arrays can be declared as outlier array(s). I > have tried but failed to find any explanation. > > Your reply (replies) would be very helpful. > > Kind regards, > > Chintanu > > [[alternative HTML version deleted]] > > > > > _______________________________________________ > Bioconductor mailing list > Bioconductor at stat.math.ethz.ch > https://stat.ethz.ch/mailman/listinfo/bioconductor > Search the archives: http://news.gmane.org/gmane.science.biology.informatics.conductor -- Wolfgang Huber EMBL http://www.embl.de/research/units/genome_biology/huber

Dear Wolfgang, Thank you for your reply. And, sorry for being late as apparently the CC- of the replied Email (if there was any) didn't reach my Email account. ad 1.: Thank you ! ad 2.: Allow me to put the question this way - had there been one (or, two) outlier array in Fig 3.2. of the book (http://tinyurl.com/2ervgpn), what would you expect to see in the heatmap ? Would the figure, then, look like the heatmap [of Fig. 1(d)] presented on the final page of arrayQualityMetrics-paper (while ignoring the right dendogram so that framework-wise comparison is possible with Fig 3.2])? However, in Fig. 1(d), array 1 looks to be an outlier rather than array 18, as mentioned in that figure ! greetings, Chintanu = = = == = = = = = = Dear Chintanu ad 1.: the assumption is that the arrays that you compare study similar biological system, e.g. the same cell line treated with different drugs, such that the gene expression differences for most genes are relatively minor. The assumption would not hold if you compare, say, a cow liver to human glioblastoma. ad 2.: I am not sure I understand your question, and maybe you need to precisify it. The heatmap is a data exploration tool, there is no single right way to read it. For instance, in Fig 3.2. that you pointed to, I'd say there are three clusters (batches) of arrays, with two of them a bit more similar to each other than to the third. Best wishes Wolfgang [[alternative HTML version deleted]]