Applying the linear models for microarray data (Limma model) [ 25], 2179 and 3193 genes were identified as responsive to A. euteiches at 6 hpi and 20 hpi, respectively, while 1610 and 1826 genes were found as responsive to P. pisi at 6 hpi and 20 hpi, respectively (P ≤ 0.05) (Table 1, Additional files 2 and 3).
In this article, we borrow the idea from the Positional Dependent Nearest Neighborhood (PDNN) model, originally developed for analyzing microarray data, to model the non-uniformity of read distribution in RNA-seq data.
This study used Bayesian GRN technology and microarray data to model the regulatory interactions after serum factor deprivation of EC, which induces cell cycle arrest and apoptosis.
To objectively select those secondary data sources that performed well for the training microarray data sets, models built on an individual secondary data source with a better average performance than the baseline model (without inclusion of secondary data) were given a decreasing score starting from 15 for the best model.
The step A models are represented as M (model based on microarray data) and G (model based on genomics data), and the step C model based on both microarray and genomics data as MG.
Regarding time-course microarray data, several model-based clustering methods have been proposed (Luan and Li, 2003; Ramoni et al., 2002; Wu et al., 2005).
Comparing with the previous models that only employ microarray data, the proposed model can bridge the gap between the relative background frequency of the observed nucleotide and the gene's transcription rate.
The step A models are MT0 (model based on microarray data at T0), MT1 (model based on microarray data at T1), PT0 (model based on proteomics data at T0), and PT1 (model based on proteomics data at T1).
The availability of both genomic and microarray data in several model organisms has opened up possibilities to elucidate important genetic regulatory mechanisms in these models.
