OVERVIEW

ASCA-genes [2] is an adaptation of the ASCA method (ANOVA Simultaneous Component Analysis) [3] developed by Smilde and co-workers, to the analysis of multifactorial experiments in transcriptomics. Basically, ASCA uses ANOVA to decompose data variation associated to experimental factors, and PCA to discover principal patterns of variation associated to the experimental factors. ASCA-genes combines this multivariate descriptive analysis on time course expression data with a gene selection procedure. ASCA-genes has been implemented for designed experiments comprising either one, two or three experimental factors, one of them typically the time. The program returns trajectory charts representing major transcriptional changes and lists of selected genes that significantly follow these major changes. An additional list collects genes with expression profile changes different from the major trends.

Parameters:

• Data: txt file with expression data, genes in rows, arrays in columns. The file must contain an additional row with arrays names and a column with gene names.
  

Example:

• Covariates: txt file with experimental design information, containing as many columns as arrays and as many rows as experimental factor. Each cell contains the value of the array in the experimental factor.
  

Example:

• Include interaction between factors: logical to indicate whether interaction/s between factors should be analyzed (only for experimental designs with more than one factor).
  
• Join interaction with second factor: logical to indicate whether interaction/s must be analysed jointly with the second/third factor (TRUE, default) or not (FALSE).
  
• Variability threshold for component selection: Criterion for selection of principal components. With “average”, components are selected that explain more than the average component variability, calculated as total data variability divided by the rank of the matrix associated to the factor. Also a fixed value for percentage of explained variability can be indicated such as 0.2, 0.4, 0.5, 0.7, 0.8…
  
• Criterion for gene selection. Strategy for selection of genes based on SPE and leverage values. Can be either
  • Resampling: where a null Leverage distribution is created by permuting columns of expression data and genes are selected at the “alpha” percentile of the null distribution and SPE cutoff is computed by using an approximation to a weighted chi-squared distribution [2].
  • minAS: where a density function is calculated on the data and genes are selected on a local minimum basis [7].
  • Gamma: where a gamma distribution is adjusted to the distributions of SPE and leverage values, and genes are selected at the “alpha” percentile of the gamma distribution [7].