How does the rank product calculate the fold change
2
0
Entering edit mode
@hildebrand-georg-3940
Last seen 9.6 years ago
Hi, my question is accordin to the library(RankProd) package maybe this question is very easy, but i was not able to find an satisfying answer. I created a list of diff. expressed genes with the Rank product package. Everything works well, but in the output of the topgenes() function i do not understand how the fold change is processed. Has anyone a hint? It is not just the difference of the mean like in limma. Even in the papers of breitling there is no good hint. best regards?, Georg Hildebrand -------------------------------- Contact/Kontakt: -------------------------------- Abteilung/Dept.:11100ho Web: http://circadianrhythms.mpibpc.mpg.de/ E-mail: georg.hildebrand at mpibpc.mpg.de
limma limma • 2.2k views
ADD COMMENT •
0
Entering edit mode
@georg-hildebrand-3943
Last seen 9.6 years ago
Hi, my question is accordin to the library(RankProd) package maybe this question is very easy, but i was not able to find an satisfying answer. I created a list of diff. expressed genes with the Rank product package. Everything works well, but in the output of the topgenes() function i do not understand how the fold change is processed. Has anyone a hint? It is not just the difference of the mean like in limma. Even in the papers of breitling there is no good hint. best regards Georg -------------- fyi here is my code. I have 3 treat. and 3 control samples for one timepoint >cgroup = 28:30 >srgroup = 40:42 >cols = c(srgroup,cgroup) # checked >data.rp = data.rma[,cols] #checked (log 2 expr. matrix) #control vs treatment class >cl = rep(c(0,1),c(3,3))#vector to classify the data1 is case 0 is >origin = rep(1, ncol(data.rp)) #data from same origin >gene.names = rownames(data.rp) >RP.out <- RPadvance(data.rp, cl, origin, num.perm = 100, + logged = TRUE, na.rm = FALSE, gene.names = gene.names, plot = TRUE, + rand = 123) >RP.out <- RP(data.rp, cl, num.perm = 100, + logged = TRUE, na.rm = FALSE, gene.names = gene.names, plot = TRUE, + rand = 123) >top.genes = topGene(RP.out, cutoff = 0.05, method = "pfp", logged = TRUE, + logbase = 2, gene.names = gene.names)
ADD COMMENT •
0
Entering edit mode
@juan-c-oliveros-collazos-2665
Last seen 9.6 years ago
Dear Georg, Please check if your Fold Change values correspond to: FC = 2**logRatio (when logRatio >= 0) and FC = -2**.(-logRatio) (when logRatio < 0) logRatio is the difference of means assuming that the means are log(2) intensities. This is the usual way to define fold changes as an intuitive measure of the level of induction / repression of a gene in a symmetric scale (with the same absolute values for induced genes and repressed genes). Best, Juan Carlos Quoting "Hildebrand, Georg" <georg.hildebrand@mpibpc.mpg.de>: > Hi, > > my question is accordin to the library(RankProd) package > maybe this question is very easy, but i was not able to find an > satisfying answer. > > I created a list of diff. expressed genes with the Rank product package. > Everything works well, but in the output of the topgenes() function i > do not understand how the fold change is processed. > Has anyone a hint? > > It is not just the difference of the mean like in limma. Even in the > papers of breitling there is no good hint. > > > best regards , > Georg Hildebrand > -------------------------------- > Contact/Kontakt: > -------------------------------- > Abteilung/Dept.:11100ho > Web: http://circadianrhythms.mpibpc.mpg.de/ > E-mail: georg.hildebrand@mpibpc.mpg.de > > _______________________________________________ > Bioconductor mailing list > Bioconductor@stat.math.ethz.ch > https://stat.ethz.ch/mailman/listinfo/bioconductor > Search the archives: > http://news.gmane.org/gmane.science.biology.informatics.conductor > [[alternative HTML version deleted]]
ADD COMMENT •
0
Entering edit mode
Hi Juan, Thanks for that help. The easies answers are mostly the best ones J. Georg From: Juan Carlos Oliveros [mailto:oliveros@cnb.csic.es] Sent: Sonntag, 21. Februar 2010 13:31 To: Hildebrand, Georg Cc: bioconductor@stat.math.ethz.ch Subject: Re: [BioC] How does the rank product calculate the fold change Dear Georg, Please check if your Fold Change values correspond to: FC = 2**logRatio (when logRatio >= 0) and FC = -2**.(-logRatio) (when logRatio < 0) logRatio is the difference of means assuming that the means are log(2) intensities. This is the usual way to define fold changes as an intuitive measure of the level of induction / repression of a gene in a symmetric scale (with the same absolute values for induced genes and repressed genes). Best, Juan Carlos Quoting "Hildebrand, Georg" <georg.hildebrand@mpibpc.mpg.de>: > Hi, > > my question is accordin to the library(RankProd) package > maybe this question is very easy, but i was not able to find an > satisfying answer. > > I created a list of diff. expressed genes with the Rank product package. > Everything works well, but in the output of the topgenes() function i > do not understand how the fold change is processed. > Has anyone a hint? > > It is not just the difference of the mean like in limma. Even in the > papers of breitling there is no good hint. > > > best regards , > Georg Hildebrand > -------------------------------- > Contact/Kontakt: > -------------------------------- > Abteilung/Dept.:11100ho > Web: http://circadianrhythms.mpibpc.mpg.de/ > E-mail: georg.hildebrand@mpibpc.mpg.de > > _______________________________________________ > Bioconductor mailing list > Bioconductor@stat.math.ethz.ch > https://stat.ethz.ch/mailman/listinfo/bioconductor > Search the archives: > http://news.gmane.org/gmane.science.biology.informatics.conductor > [[alternative HTML version deleted]]
ADD REPLY •
0
Entering edit mode
Hi This is a great opportunity to toot about the advantages of open source and good programming habits: If you just type "RP" in the R prompt, you will get the code that is being executed. If you browse it, you will see the following lines which clearly tells you how the fold change is being calculated in RankProd. **************************************************** if (!is.null(y)) { num.class <- 2 data1 <- as.matrix(data[, x]) data2 <- as.matrix(data[, y]) data1.ave = apply(data1, 1, mean) data2.ave = apply(data2, 1, mean) if (logged) { fold.change = data1.ave - data2.ave } else { fold.change = data1.ave/data2.ave } } **************************************************** -best -Lax On Mon, Feb 22, 2010 at 3:29 AM, Hildebrand, Georg <georg.hildebrand at="" mpibpc.mpg.de=""> wrote: > Hi Juan, > > > > Thanks for that help. The easies answers are mostly the best ones J. > > Georg > > > > > > From: Juan Carlos Oliveros [mailto:oliveros at cnb.csic.es] > Sent: Sonntag, 21. Februar 2010 13:31 > To: Hildebrand, Georg > Cc: bioconductor at stat.math.ethz.ch > Subject: Re: [BioC] How does the rank product calculate the fold change > > > > Dear Georg, > > Please check if your Fold Change values correspond to: > > FC = 2**logRatio (when logRatio >= 0) > > and > > FC = -2**.(-logRatio) (when logRatio < 0) > > logRatio is the difference of means assuming that the means are log(2) > intensities. > > This is the usual way to define fold changes as an intuitive measure of the > level of induction / repression of a gene in a symmetric scale (with the same > absolute values for induced genes and repressed genes). > > Best, > > Juan Carlos > > Quoting "Hildebrand, Georg" <georg.hildebrand at="" mpibpc.mpg.de="">: > >> Hi, >> >> my question is accordin to the library(RankProd) package >> maybe this question is very easy, but i was not able to find an >> satisfying answer. >> >> I created a list of diff. expressed genes with the Rank product package. >> Everything works well, but in the output of the topgenes() function i >> do not understand how the fold change is processed. >> Has anyone a hint? >> >> It is not just the difference of the mean like in limma. Even in the >> papers of breitling there is no good hint. >> >> >> best regards , >> Georg Hildebrand >> -------------------------------- >> Contact/Kontakt: >> -------------------------------- >> Abteilung/Dept.:11100ho >> Web: http://circadianrhythms.mpibpc.mpg.de/ >> E-mail: georg.hildebrand at mpibpc.mpg.de >> >> _______________________________________________ >> 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 >> > > > > > ? ? ? ?[[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 >
ADD REPLY •

Login before adding your answer.

Similar Posts
Loading Similar Posts
Traffic: 729 users visited in the last hour
Content Search
Users
Tags
Badges
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