Multi-Pipeline Gene Expression Analysis

The Multi-Pipeline Gene Expression Analysis tool was developed as a prototype during the Bioinformatics Integration Support Contract (BISC) in 2010-2011. It was available online at the ImmPort website in an "Alpha" release until November 2013.

The tool utilizes about 300 gene expression analysis pipelines with combinations of methods for:
1. Background Correction (none, mas5, rma, DFCM, RMA-mean, RMA-75)
2. Normalization (none, constant, contrasts, quantiles, invariant set, loess)
3. Summarization (mas5, rma, farms, DFW)
4. Gene Selection (SAM, ANOVA)

Then executes K-means clustering on the significant genes, and evaluates the pipelines using the cumulative distribution funciton of the GO term co-clustering p-values. From this, the optimal microarray data workflow is chosen.

Input files are CEL files along with a CLM file, which is a tab delimited text file containing one scan, sample and class per line.