Question: query on variance parameters for RMA
0
gravatar for Hugh Shanahan
6.7 years ago by
Hugh Shanahan50 wrote:
Hi, we would like to compute the parameters (i.e, the other parameters that are computed via the median polish algorithm) that are computed in the final summarisation step in RMA. In particular, we'd like to determine the parameter alpha_i that tells us how the estimate for the expression varies as a function of the probe. I guess it's possible to open up the relevant C-code and find it there, but are there are libraries that allow one to do it more easily ? I saw that affyPLM appears to be doing something similar - is this where we should be looking ? Many thanks, Hugh
probe • 445 views
ADD COMMENTlink modified 6.7 years ago by James W. MacDonald51k • written 6.7 years ago by Hugh Shanahan50
Answer: query on variance parameters for RMA
0
gravatar for James W. MacDonald
6.7 years ago by
United States
James W. MacDonald51k wrote:
Hi Hugh, It's simple enough to get these data on a probeset-by-probeset basis: > library(affydata) > data(Dilution) > medpolish(log2(pm(Dilution, "1007_s_at"))) 1: 5.536431 2: 4.554366 Final: 4.528792 Median Polish Results (Dataset: "log2(pm(Dilution, "1007_s_at"))") Overall: 8.618676 Row Effects: 1007_s_at1 1007_s_at2 1007_s_at3 1007_s_at4 1007_s_at5 1007_s_at6 -0.59973136 0.29592898 0.78257437 2.05732074 2.39466709 1.73948500 1007_s_at7 1007_s_at8 1007_s_at9 1007_s_at10 1007_s_at11 1007_s_at12 -0.04027435 -0.06544218 -0.67467511 0.69974913 0.51478816 -0.13604974 1007_s_at13 1007_s_at14 1007_s_at15 1007_s_at16 -0.15326752 -1.90470573 -1.11026444 0.04027435 Column Effects: 20A 20B 10A 10B 0.5728390 -0.1144511 0.1202443 -0.6362839 Residuals: 20A 20B 10A 10B 1007_s_at1 -0.18749394 -0.0465129 0.0407198 0.17422802 1007_s_at2 0.13826450 0.0295685 -0.1240571 -0.03039400 1007_s_at3 0.07439714 -0.0317711 0.0259779 -0.03874874 1007_s_at4 -0.03633972 0.1518405 -0.0181643 0.01217539 1007_s_at5 0.07528392 -0.0683772 0.0625840 -0.09050186 1007_s_at6 0.00379623 0.0646288 -0.0087639 -0.11399244 1007_s_at7 0.04055849 -0.0460984 -0.0440103 0.03802140 1007_s_at8 -0.11484593 0.0062318 0.0840782 -0.00705727 1007_s_at9 -0.02900022 -0.0267107 0.1319877 0.02588521 1007_s_at10 -0.12805213 0.1904883 0.0268966 -0.03288549 1007_s_at11 -0.00067114 0.0083397 -0.1167177 0.00067114 1007_s_at12 0.02934276 -0.0062318 -0.2409272 0.00540625 1007_s_at13 0.00067114 0.0109860 -0.1899048 -0.00067114 1007_s_at14 -0.09698507 -0.0145570 0.0087639 0.07814337 1007_s_at15 -0.26554742 -0.0828937 0.1806296 0.08206822 1007_s_at16 0.03967331 0.1902100 -0.2177162 -0.03967331 And doing it on a whole array isn't that expensive: > pms <- pm(Dilution, LISTRUE = TRUE) > system.time(meds <- lapply(pms, medpolish, trace.iter=FALSE)) user system elapsed 62.387 0.033 62.439 Best, Jim On 2/20/2013 7:39 AM, Hugh Shanahan wrote: > Hi, > we would like to compute the parameters (i.e, the other parameters that are computed via the median polish algorithm) that are computed in the final summarisation step in RMA. In particular, we'd like to determine the parameter alpha_i that tells us how the estimate for the expression varies as a function of the probe. I guess it's possible to open up the relevant C-code and find it there, but are there are libraries that allow one to do it more easily ? I saw that affyPLM appears to be doing something similar - is this where we should be looking ? > > Many thanks, > Hugh > > _______________________________________________ > Bioconductor mailing list > Bioconductor at r-project.org > https://stat.ethz.ch/mailman/listinfo/bioconductor > Search the archives: http://news.gmane.org/gmane.science.biology.informatics.conductor -- James W. MacDonald, M.S. Biostatistician University of Washington Environmental and Occupational Health Sciences 4225 Roosevelt Way NE, # 100 Seattle WA 98105-6099
ADD COMMENTlink written 6.7 years ago by James W. MacDonald51k
I should note that ideally you would first background correct and normalize and then do what I showed below: Dilution <- bg.correct(Dilution, "rma") Dilution <- normalize(Dilution) Best, Jim On 2/20/2013 11:07 AM, James W. MacDonald wrote: > Hi Hugh, > > It's simple enough to get these data on a probeset-by-probeset basis: > > > library(affydata) > > data(Dilution) > > medpolish(log2(pm(Dilution, "1007_s_at"))) > 1: 5.536431 > 2: 4.554366 > Final: 4.528792 > > Median Polish Results (Dataset: "log2(pm(Dilution, "1007_s_at"))") > > Overall: 8.618676 > > Row Effects: > 1007_s_at1 1007_s_at2 1007_s_at3 1007_s_at4 1007_s_at5 1007_s_at6 > -0.59973136 0.29592898 0.78257437 2.05732074 2.39466709 1.73948500 > 1007_s_at7 1007_s_at8 1007_s_at9 1007_s_at10 1007_s_at11 1007_s_at12 > -0.04027435 -0.06544218 -0.67467511 0.69974913 0.51478816 -0.13604974 > 1007_s_at13 1007_s_at14 1007_s_at15 1007_s_at16 > -0.15326752 -1.90470573 -1.11026444 0.04027435 > > Column Effects: > 20A 20B 10A 10B > 0.5728390 -0.1144511 0.1202443 -0.6362839 > > Residuals: > 20A 20B 10A 10B > 1007_s_at1 -0.18749394 -0.0465129 0.0407198 0.17422802 > 1007_s_at2 0.13826450 0.0295685 -0.1240571 -0.03039400 > 1007_s_at3 0.07439714 -0.0317711 0.0259779 -0.03874874 > 1007_s_at4 -0.03633972 0.1518405 -0.0181643 0.01217539 > 1007_s_at5 0.07528392 -0.0683772 0.0625840 -0.09050186 > 1007_s_at6 0.00379623 0.0646288 -0.0087639 -0.11399244 > 1007_s_at7 0.04055849 -0.0460984 -0.0440103 0.03802140 > 1007_s_at8 -0.11484593 0.0062318 0.0840782 -0.00705727 > 1007_s_at9 -0.02900022 -0.0267107 0.1319877 0.02588521 > 1007_s_at10 -0.12805213 0.1904883 0.0268966 -0.03288549 > 1007_s_at11 -0.00067114 0.0083397 -0.1167177 0.00067114 > 1007_s_at12 0.02934276 -0.0062318 -0.2409272 0.00540625 > 1007_s_at13 0.00067114 0.0109860 -0.1899048 -0.00067114 > 1007_s_at14 -0.09698507 -0.0145570 0.0087639 0.07814337 > 1007_s_at15 -0.26554742 -0.0828937 0.1806296 0.08206822 > 1007_s_at16 0.03967331 0.1902100 -0.2177162 -0.03967331 > > And doing it on a whole array isn't that expensive: > > > pms <- pm(Dilution, LISTRUE = TRUE) > > system.time(meds <- lapply(pms, medpolish, trace.iter=FALSE)) > user system elapsed > 62.387 0.033 62.439 > > Best, > > Jim > > > > On 2/20/2013 7:39 AM, Hugh Shanahan wrote: >> Hi, >> we would like to compute the parameters (i.e, the other >> parameters that are computed via the median polish algorithm) that >> are computed in the final summarisation step in RMA. In particular, >> we'd like to determine the parameter alpha_i that tells us how the >> estimate for the expression varies as a function of the probe. I >> guess it's possible to open up the relevant C-code and find it there, >> but are there are libraries that allow one to do it more easily ? I >> saw that affyPLM appears to be doing something similar - is this >> where we should be looking ? >> >> Many thanks, >> Hugh >> >> _______________________________________________ >> Bioconductor mailing list >> Bioconductor at r-project.org >> https://stat.ethz.ch/mailman/listinfo/bioconductor >> Search the archives: >> http://news.gmane.org/gmane.science.biology.informatics.conductor > -- James W. MacDonald, M.S. Biostatistician University of Washington Environmental and Occupational Health Sciences 4225 Roosevelt Way NE, # 100 Seattle WA 98105-6099
ADD REPLYlink written 6.7 years ago by James W. MacDonald51k
Please log in to add an answer.

Help
Access

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 16.09
Traffic: 151 users visited in the last hour