Newbe question
1
0
Entering edit mode
SimonNoël ▴ 450
@simonnoel-3455
Last seen 7.7 years ago
Hi, I have some kind of a shameful question... My director say that I will need to do a logistic regression in order to test for gene- gene interaction... But what's that? I tryed to search for a package that do logistic regression on bioconductor and I get around 40 hit... Witch one should I use? Simon No??l CdeC
Regression Regression • 454 views
ADD COMMENT
0
Entering edit mode
SimonNoël ▴ 450
@simonnoel-3455
Last seen 7.7 years ago
Hi, I have some kind of a shameful question... My director say that I will need to do a logistic regression in order to test for gene-gene interaction... But what's that? I tryed to search for a package that do logistic regression on bioconductor and I get around 40 hit... Witch one should I use? Simon No??l CdeC
ADD COMMENT
0
Entering edit mode
Please recognize that the following is by no means authoritative. That said, nobody else was providing an answer (perhaps for good reason!), so... A typical workflow (if you've got candidate loci, say, 'gene1' and 'gene2' calls, along with some other genotyped loci that you don't particularly care about): # y is a binary response, e.g. unaffected=0, affected=1 > y = your.outcomes # we assume there will usually be population stratification; # attempt to control for it using principal components of X > library(pcaMethods) > PCs = nipalsPca(t(your.data), nPcs=3) # can tolerate NAs # I'm happy if max(R2cum) > 0.8; if not, set nPcs higher > PCs@R2cum [1] 0.8977354 0.9925331 1.0000000 > rsv = PCs@scores > colnames(rsv) = paste('PC', 1:3, sep='') > your.data = cbind(y, your.data, rsv) # fit a logistic model w/main effects and an interaction, # controlling (somewhat) for stratification (...maybe) > fit = with(your.data, glm( y ~ gene1*gene2 + PC1 + PC2 + PC3, family='binomial')) # see what it tells you > summary(fit) # nb. if you had parent-child trios, you could instead use # conditional logistic regression to emulate a TDT or FBAT Hopefully you have candidate loci, because otherwise the simplistic exercise above will be replaced by an entirely different kettle of fish if you don't. But then you should work with someone who does this stuff for a living and not just randomly solicited, possibly untrustworthy advice from the Interwebs. IMHO, of course. --t 2011/7/23 Simon Noël <simon.noel.2@ulaval.ca> > > Hi, > > > > I have some kind of a shameful question... My director say that I will > need > to do a logistic regression in order to test for gene-gene interaction... > But what's that? I tryed to search for a package that > do logistic regression on bioconductor and I get around 40 hit... Witch > one should I use? > > > Simon Noël > CdeC > _______________________________________________ > Bioconductor mailing list > Bioconductor@r-project.org > https://stat.ethz.ch/mailman/listinfo/bioconductor > Search the archives: > http://news.gmane.org/gmane.science.biology.informatics.conductor > -- If people do not believe that mathematics is simple, it is only because they do not realize how complicated life is. John von Neumann<http: www-groups.dcs.st-="" and.ac.uk="" ~history="" biographies="" von_neumann.html=""> [[alternative HTML version deleted]]
ADD REPLY

Login before adding your answer.

Traffic: 286 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