My targets was Sex Patient Treatment M 1 N M 1 T M 2 N M 2 T M 3 N M 3 T M 4 N M 4 T M 5 N M 5 T F 6 N F 6 T F 7 N F 7 T F 8 N F 8 T F 9 N F 9 T Thanks for your response. -------------- Huajing Teng Beijing Institutes of Life Science, Chinese Academy of Sciences Beichen West Road,Chao Yang District,Beijing 100101, China E-mail:tenghj at mail.biols.ac.cn 2013-01-09

Looking at the result of running that table through the code that you provided in your first email, I see that the resulting design matrix has 12 columns but only a rank of 11 (as measured by "qr(design)$rank"). I think this may be the problem. I have observed this behavior from model.matrix before with complex model formulae like the one you are using, but I'm not sure how to fix it. -Ryan On Tue 08 Jan 2013 04:59:12 PM PST, Huajing Teng wrote: > My targets was > Sex Patient Treatment > M 1 N > M 1 T > M 2 N > M 2 T > M 3 N > M 3 T > M 4 N > M 4 T > M 5 N > M 5 T > F 6 N > F 6 T > F 7 N > F 7 T > F 8 N > F 8 T > F 9 N > F 9 T > Thanks for your response. > > -------------- > Huajing Teng > Beijing Institutes of Life Science, Chinese Academy of Sciences > Beichen West Road,Chao Yang District,Beijing 100101, China > E-mail:tenghj at mail.biols.ac.cn > 2013-01-09 > > _______________________________________________ > 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