Page History
...
In GSA, the data are used to choose the best model for each transcript. Then, the best model produces p-values, fold change estimates, and other measures of importance for the factor(s) of interest.
Figure 1:
Numbered figure captions | ||||
---|---|---|---|---|
| ||||
Currently, GSA is capable of considering the following five response distributions: Normal, Lognormal, Lognormal with shrinkage, Negative Binomial, Poisson (Figure 1). The GSA interface has an option to restrict this pool of distributions in any way, e.g. by specifying just one response distribution. The user also specifies the factors that may enter the model (Figure 2A) and comparisons for categorical factors (Figure 2B).
Figure 2A:
Numbered figure captions | ||||
---|---|---|---|---|
| ||||
Figure 2B: Choosing comparisons in GSA
...
Overview
Content Tools