Dear Benilton, Naomi, Ben, and Monnie, Thank you for your answers and references. I know understand RMA well enough to explain it simply to my class and to the experimentalists for whom I provide bioinformatics support I had misunderstood a critical aspect of the algorithm previously. The point you cleared up also has consequences for my understanding of GCRMA, so that I will reread those papers and probably have questions on that as well. One quick question: To what do the boundaries of the boxes in the affy boxplots refer? Quartiles? Standard deviations? Other? Thanks and best wishes, Rich On Jan 30, 2006, at 6:48 PM, Benilton Carvalho wrote: > Dear Dr. Friedman, > > I'd say that what Rafa wants to say in his paper is: > > The PMs are a combination of background noise and signal: > (PM.inj = bg.inj + s.ijn) > > We're not interested in this noisy information. We really want the > signal s.ijn. But since we can only observe the noisy information > (PM), we have to estimate the signal. > > Supposed you observed PM = 5,827.82. Then the formulae below answer > the following question: what's the best value for the true signal if > the noisy observation is 5,827.82? > > The usual answer for the is: the average... E(sijn | PMinj) > > The hypothetical thing here would be: consider a situation where you > always know the true signal s.ijn. Now, write down all the signals > s.ijn_k where PM = 5,827.82 and take the mean of such s.ijn_k. > > Of course we can not have such situation in real life. For this > reason, if one assumes distributions (exponential for the signal and > normal for the background - remember PM = signal + background), we can > simulate the real signal and the real background in order to get the > noisy observation (PM) and then we're good to use the hypothetical > situation above. > > Such simulation would consume a good portion of time, but the > distribuitions above (normal and exponential) allow us to derive the > exact answer without the need of simulation. > > I'm sure this is not the best way to explain, but I hope it helps. > > kind regards, > > benilton > > > > >> Naomi, >> >> I am trying to understand RMA better myself at the moment. I cannot >> understand how >> the background is obtained from page 6 of the pdf file that you cite >> below. In >> the RMA paper (Biostat 4, 249-264). Dr. Irizarry and coworkers state >> "An alternative >> background correction is to consider B(PM(ijn))=E(sijn)|PMijn). ... To >> obtain a computationally >> feasible B(.) we consider a closed form transformation obtained when >> sijn >> is exponential and bgijn is normal." >> This text seems to refer to the graphs on page 6 of what you cite >> below, but I don't >> understand it. Can you (or anybody out there) please explain it in >> simpler language. >> >> Thanks and best wishes, >> Rich >> >> On Jan 30, 2006, at 5:24 PM, Naomi Altman wrote: >> >>> I found the answer to my questions on Dr. Irizarry's website in his >>> talk: >>> >>> http://www.iobion.com/slides/RMA/rma.pdf >>> >>> Naomi S. Altman 814-865-3791 (voice) >>> Associate Professor >>> Dept. of Statistics 814-863-7114 (fax) >>> Penn State University 814-865-1348 >>> (Statistics) >>> University Park, PA 16802-2111 >>> >>> _______________________________________________ >>> Bioconductor mailing list >>> Bioconductor at stat.math.ethz.ch >>> https://stat.ethz.ch/mailman/listinfo/bioconductor >>> >> ------------------------------------------------------------ >> Richard A. Friedman, PhD >> Associate Research Scientist >> Herbert Irving Comprehensive Cancer Center >> Oncoinformatics Core >> Lecturer >> Department of Biomedical Informatics >> Box 95, Room 130BB or P&S 1-420C >> Columbia University Medical Center >> 630 W. 168th St. >> New York, NY 10032 >> (212)305-6901 (5-6901) (voice) >> friedman at cancercenter.columbia.edu >> http://cancercenter.columbia.edu/~friedman/ >> >> "42 is the answer. Dylan got it wrong. 'Blowin' >> in the wind' is not the answer. It isn't even >> a number' " - Rose Friedman, age 9 >> >> _______________________________________________ >> Bioconductor mailing list >> Bioconductor at stat.math.ethz.ch >> https://stat.ethz.ch/mailman/listinfo/bioconductor >> > > > -- > Benilton Carvalho > PhD Candidate > Department of Biostatistics > Bloomberg School of Public Health > Johns Hopkins University > Baltimore, MD USA > ------------------------------------------------------------ Richard A. Friedman, PhD Associate Research Scientist Herbert Irving Comprehensive Cancer Center Oncoinformatics Core Lecturer Department of Biomedical Informatics Box 95, Room 130BB or P&S 1-420C Columbia University Medical Center 630 W. 168th St. New York, NY 10032 (212)305-6901 (5-6901) (voice) friedman at cancercenter.columbia.edu http://cancercenter.columbia.edu/~friedman/ "42 is the answer. Dylan got it wrong. 'Blowin' in the wind' is not the answer. It isn't even a number' " - Rose Friedman, age 9