Hello, A while ago I had posted a question about amplified> RNA experiment and guidelines one should follow during data analysis and Dr. Gentleman had suggested that RNA digestion plot will be a good indicator for deciding which samples should be used in data analysis as it is important to comapre likes. For my experiments,I think prima facie the plots look quite a bit similar but since I am new to this area I cannot estimate what should be range of difference in slopes that one should tolerate while choosing the comparable profiles. To give an idea of what I am seeing I am attaching a postscript image and giving the slopes at the bottom of this email. I will be extremely obliged if somebody could provide me further guidance so that I know I am on the right track. Please let me know if I need to provide more than the following information? Thanks, Swati Here is the data: (I think sample # 8 seems to have a slope that is different fro the rest!! Should it be eliminated or one should observe the trend of the graph?) $slope [1] 1.6329065 1.5580935 1.5640235 0.9613637 1.2156617 1.2004132 1.4831510 [8] 0.3772531 1.6212330 1.2952673 1.2148458 1.6020575 1.8564214 1.0944649 __________________________________________________ -------------- next part -------------- A non-text attachment was scrubbed... Name: RNAdeg.ps Type: application/postscript Size: 6836 bytes Desc: RNAdeg.ps Url : https://stat.ethz.ch/pipermail/bioconductor/attachments/20050328 /00f4a0c1/RNAdeg.ps

Looking at the plot, I would not take the slopes too seriously! I only eliminate an array if it appears to be dramatically different from the others. Your arrays vary (which is expected) but I do not see any outlying array. I suppose you could do a formal test, but I wouldn't bother. To do the test, you would probably want to use a semiparametric model of the form curve(i) = m(i) + s(x) where s(x) is the mean curve and m(i) is a random array effect. Once you fit the model, you could test the goodness of fit of each curve to the model, adjust for multiple comparisons, and reject any significant curves. Sorry, I do not have a reference for this. --Naomi At 08:35 AM 3/28/2005, Swati Ranade wrote: > Hello, > > A while ago I had posted a question about amplified> >RNA experiment and guidelines one should follow >during data analysis and Dr. Gentleman had suggested >that RNA digestion plot will be a good indicator for >deciding which samples should be used in data analysis >as it is important to comapre likes. For my >experiments,I think prima facie the plots look quite a >bit similar but since I am new to this area >I cannot estimate what should be range of difference >in slopes that one should tolerate while choosing the >comparable profiles. To give an idea of what I am >seeing I am attaching a postscript image and giving >the slopes at the bottom of this email. I will be >extremely obliged if somebody could provide me >further guidance so that I know I am on the right >track. Please let me know if I need to provide more >than the following information? > > Thanks, > > Swati > >Here is the data: (I think sample # 8 seems to have >a slope that is different fro the rest!! Should it be >eliminated or one should observe the trend of the >graph?) >$slope > [1] 1.6329065 1.5580935 1.5640235 0.9613637 > 1.2156617 1.2004132 1.4831510 > [8] 0.3772531 1.6212330 1.2952673 1.2148458 > 1.6020575 1.8564214 1.0944649 > > > > >__________________________________________________ > > > >_______________________________________________ >Bioconductor mailing list >Bioconductor@stat.math.ethz.ch >https://stat.ethz.ch/mailman/listinfo/bioconductor Naomi S. Altman 814-865-3791 (voice) Associate Professor Bioinformatics Consulting Center Dept. of Statistics 814-863-7114 (fax) Penn State University 814-865-1348 (Statistics) University Park, PA 16802-2111

Thanks Naomi. Will try your suggestion, but I get the point, the trends of the curves are more important than the slopes alone. --- Naomi Altman <naomi@stat.psu.edu> wrote: > Looking at the plot, I would not take the slopes too > seriously! > > I only eliminate an array if it appears to be > dramatically different from > the others. Your arrays vary (which is expected) > but I do not see any > outlying array. > > I suppose you could do a formal test, but I wouldn't > bother. To do the > test, you would probably want to use a > semiparametric model of the form > > curve(i) = m(i) + s(x) > > where s(x) is the mean curve and m(i) is a random > array effect. Once you > fit the model, you could test the goodness of fit of > each curve to the > model, adjust for multiple comparisons, and reject > any significant > curves. Sorry, I do not have a reference for this. > > --Naomi > > At 08:35 AM 3/28/2005, Swati Ranade wrote: > > > Hello, > > > > A while ago I had posted a question about > amplified> > >RNA experiment and guidelines one should follow > >during data analysis and Dr. Gentleman had > suggested > >that RNA digestion plot will be a good indicator > for > >deciding which samples should be used in data > analysis > >as it is important to comapre likes. For my > >experiments,I think prima facie the plots look > quite a > >bit similar but since I am new to this area > >I cannot estimate what should be range of > difference > >in slopes that one should tolerate while choosing > the > >comparable profiles. To give an idea of what I am > >seeing I am attaching a postscript image and giving > >the slopes at the bottom of this email. I will be > >extremely obliged if somebody could provide me > >further guidance so that I know I am on the right > >track. Please let me know if I need to provide more > >than the following information? > > > > Thanks, > > > > Swati > > > >Here is the data: (I think sample # 8 seems to have > >a slope that is different fro the rest!! Should it > be > >eliminated or one should observe the trend of the > >graph?) > >$slope > > [1] 1.6329065 1.5580935 1.5640235 0.9613637 > > 1.2156617 1.2004132 1.4831510 > > [8] 0.3772531 1.6212330 1.2952673 1.2148458 > > 1.6020575 1.8564214 1.0944649 > > > > > > > > > >__________________________________________________ > > > > > > > >_______________________________________________ > >Bioconductor mailing list > >Bioconductor@stat.math.ethz.ch > >https://stat.ethz.ch/mailman/listinfo/bioconductor > > Naomi S. Altman > 814-865-3791 (voice) > Associate Professor > Bioinformatics Consulting Center > Dept. of Statistics > 814-863-7114 (fax) > Penn State University > 814-865-1348 (Statistics) > University Park, PA 16802-2111 > >