This method of excluding false negative signals is a viable alternative to applying static thresholds, trying to maximize a small number of true positive hits, or applying mixed variance models that derive their power from extensive replication of both control and experimental samples.
Multivariable association between flow, treatment and other cofactors was examined in mixed variance covariance models using proc mixed from SAS.
Two submodels are presented: a simple and easy to implement Haley-Knott type regression for half-sib families, and a general mixed (variance component) model for general pedigrees.
We applied a mixed model (variance components analysis) to assess the proportion of total variation due to variation among subjects, hybridizations, labels, and chips, among the randomly selected within-subject replicates (n = 19).
We found that the random effects explained almost 10% of the mixed model variance.
At the group level, random-effects components of mixed effects variance were modelled and estimated for each contrast (FLAME, Beckmann et al., 2003).
BrdU and fusion index data were analyzed by mixed model variance analysis, using treatment (P311 and GFP control) as fixed effects with experiments and replicate as random effects.
To evaluate between group differences, a linear mixed model (variance component structure with time and time*group as fixed effects and time as random effect intercept and slope) was used (IBM® SPSS® Statistics, version 20.0 (IBM Corp., Somers, New York, USA).
As a by-product, the same results hold for multivariate linear mixed models with variances depending on treatments.
Basic concepts are overviewed and defined or re-defined: fundamental transfer modes, natural vs. controlled transfer mode, variants vs. variances, mixed vs. combined modes, drop spray transfer.
