Hello,

My name is Mahes Muniandy and I am a doctoral student working on twin data.  I have been using the WGCNA package to cluster and analyse my traits and gene expression data. Because I am using monozygotic twin data and would like to see how my twins cluster as pairs, I have taken the ratios of the values between the co-twins. In other words, I treat each pair as one sample and take twin1/twin2 to get the 'pair' values. Is this acceptable? Should I be taking delta value or doing something completely different? Incidently, I have also performed limma analysis on my expression data and I'm happy with that side of things.

Any ideas would be appreciated.

Many Thanks,

Mahes Muniandy,

Obesity Research Unit,

University of Helsinki

I'm not sure what you mean by "see how my twins cluster as pairs". If you want to treat each pair as a single unit ("pair"), you may want the average, or perhaps the geometric mean, of the expression values within each pair, then cluster the resulting meta-samples (each meta-samples represents a pair of twins). Taking the differences or ratios between siblings in a pair would work if you were to look for differences between the siblings in each pair, not for looking at differences between pairs.