The AffyComp II Benchmarking Study for the normalisation of gene chip expression measures showed that whilst RMA outperformed MAS 5.0, in certain measures GCRMA and FARMS outperformed both of them [Microarray data analysis: from disarray to consolidation and consensus D.B. Allison, X. Cui, G.P. Page and M. Sabripour Nature Reviews Geentics (2005)7 55-65].
It is possible that the appropriate method for normalisation will be data-set and biological question dependent as in some cases expression will be much stronger than in other and in some cases differences will be small and so more sensitive methods are needed at the risk of losing specificity.
Luckily both FARMS and GCRMA are both available as packages for R and so it is relatively easy to process the arrays using all of the potential methods and these normalised values can then be taken forward to the subsequent analysis where their different effects can be evaluated.
It is possible that the appropriate method for normalisation will be data-set and biological question dependent as in some cases expression will be much stronger than in other and in some cases differences will be small and so more sensitive methods are needed at the risk of losing specificity.
Luckily both FARMS and GCRMA are both available as packages for R and so it is relatively easy to process the arrays using all of the potential methods and these normalised values can then be taken forward to the subsequent analysis where their different effects can be evaluated.
No comments:
Post a Comment