Limma and Dye Flips
1
0
Entering edit mode
Sally ▴ 250
@sally-2430
Last seen 9.6 years ago
An embedded and charset-unspecified text was scrubbed... Name: not available Url: https://stat.ethz.ch/pipermail/bioconductor/attachments/20080123/ 1a12eb5a/attachment.pl
• 728 views
ADD COMMENT
0
Entering edit mode
Claus Mayer ▴ 340
@claus-mayer-1179
Last seen 9.5 years ago
European Union
Hi Sally, What you have done (averaging across the dye-flips) is basically ok, as the information you loose by doing that only concerns technical variability. The more sophisticated way of doing things would be to include all arrays, but to change your limma analysis by a) adding a dye effect (cf. limma tutorial) b) adding a block effect that informs the analysis about which arrays are technical replicates (dye-flips), that analyse the same biological samples (again see the limma tutorial for details). In addition to your treatment effect this analysis will give you information about gene-specific dye-effects and the correlation between the two dye-flips. Both of these might not be relevant to the biological question of interest, but be helpful from a quality control perspective. If you average the dye-flips you eliminate the gene-specific dye- effect (which is good), but you also loose the information about it. Equally you give up the information about the technical array to array (+dye flip)-variation. The fold-changes you estimate for your biological effect of interest will be identical for both approaches but p-values etc will differ slightly. From the point of statistical power both approaches should be similar as far as I can see (and others may correct me if I overlooked any subtle points here). Best Wishes Claus Sally wrote: > My question is about handling dye flips in Limma. > > In my experiment I did a dye flip for each comparison. When I finished the preprocessing I averaged the M values of the two dye flip arrays.Now I am fitting a linear model. Should I have averaged the dye flips before, or should the dye flips be incorporated in the linear model? > > Sally Goldes > [[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 > > > > > > > Click link below to report this email as spam. > https://www.mailcontrol.com/sr/ubMazK5f8H7cohtw78djyfAVCukN02RG8MraH iLp2brog64sJZMbax70VQcaVVh5T7KYBPDklBRfAchUXJwv+Q== > > > -- ********************************************************************** ************* Dr Claus-D. Mayer | http://www.bioss.ac.uk Biomathematics & Statistics Scotland | email: claus at bioss.ac.uk Rowett Research Institute | Telephone: +44 (0) 1224 716652 Aberdeen AB21 9SB, Scotland, UK. | Fax: +44 (0) 1224 715349
ADD COMMENT
0
Entering edit mode
A related question: Does limma algorithms for dye swap use "weights" for dye flip arrays if not every sample had dye swap replication? For example there are 3 normal vs. tumor arrays, but only one array with different dye-orientation. thanks, JJ --On Thursday, January 24, 2008 4:08 PM +0000 Claus-Dieter Mayer <claus at="" bioss.ac.uk=""> wrote: > Hi Sally, > > What you have done (averaging across the dye-flips) is basically ok, as > the information you loose by doing that only concerns technical > variability. The more sophisticated way of doing things would be to > include all arrays, but to change your limma analysis by > a) adding a dye effect (cf. limma tutorial) > b) adding a block effect that informs the analysis about which arrays > are technical replicates (dye-flips), that analyse the same biological > samples (again see the limma tutorial for details). > > In addition to your treatment effect this analysis will give you > information about gene-specific dye-effects and the correlation between > the two dye-flips. Both of these might not be relevant to the biological > question of interest, but be helpful from a quality control perspective. > If you average the dye-flips you eliminate the gene-specific dye- effect > (which is good), but you also loose the information about it. Equally > you give up the information about the technical array to array (+dye > flip)-variation. The fold-changes you estimate for your biological > effect of interest will be identical for both approaches but p-values > etc will differ slightly. From the point of statistical power both > approaches should be similar as far as I can see (and others may correct > me if I overlooked any subtle points here). > > Best Wishes > > Claus > > > > Sally wrote: >> My question is about handling dye flips in Limma. >> >> In my experiment I did a dye flip for each comparison. When I finished >> the preprocessing I averaged the M values of the two dye flip arrays.Now >> I am fitting a linear model. Should I have averaged the dye flips >> before, or should the dye flips be incorporated in the linear model? >> >> Sally Goldes >> [[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 >> >> >> >> >> >> >> Click link below to report this email as spam. >> https://www.mailcontrol.com/sr/ubMazK5f8H7cohtw78djyfAVCukN02RG8Mra HiLp2 >> brog64sJZMbax70VQcaVVh5T7KYBPDklBRfAchUXJwv+Q== >> >> >> > > -- > ******************************************************************** ***** > ********** Dr Claus-D. Mayer | http://www.bioss.ac.uk > Biomathematics & Statistics Scotland | email: claus at bioss.ac.uk > Rowett Research Institute | Telephone: +44 (0) 1224 716652 > Aberdeen AB21 9SB, Scotland, UK. | Fax: +44 (0) 1224 715349 > > _______________________________________________ > 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 ################################## Jianping Jin Ph.D. Bioinformatics scientist Center for Bioinformatics Room 3133 Bioinformatics building CB# 7104 University of Chapel Hill Chapel Hill, NC 27599 Phone: (919)843-6105 FAX: (919)843-3103 E-Mail: jjin at email.unc.edu
ADD REPLY

Login before adding your answer.

Traffic: 735 users visited in the last hour
Help About
FAQ
Access RSS
API
Stats

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.

Powered by the version 2.3.6