aroma-project.org

Aroma.affymetrix is an R package for analyzing large Affymetrix data sets.  It will let you analyze 100-1,000s of arrays of various chip types, e.g. expression arrays, SNP chips, exon arrays and so on.  Here are some of the features:

• Number of arrays: unlimited.
• System requirements: ~1.5GB RAM, any operating system.
• Chip types: all Affymetrix chip types with a CDF, e.g. gene expression, exon, SNP & CN, tiling arrays.
• Immediate support for custom CDFs, e.g. BrainArray, AffyProbeMiner, and GeneAnnot.
• File formats: Works directly with CEL and CDF files (all versions; text/ASCII, binary/XDA, binary/Calvin).
• Persistent memory: Final and intermediate results and estimates are stored on file.
• Robustness: Analysis picks up where last interrupted.
• Reproducibility: Near perfect replication of RMA and SNPRMA (normalization & summarization) as implemented in the affyPLM, affy, and oligo packages, as well as CRLMM (genotyping) of oligo.
• Portable scripts: All data files are referred to by relative pathnames.
• Export to/Import from: Bioconductor, CNAG, CNAT & dChip.
• Pre-processing: Background correction, allelic cross-talk calibration, quantile normalization, nucleotide-position normalization etc.
• Probe-level modeling: RMA (log-additive), MBEI (multiplicative), affine (multiplicative with or without offset).
• Post-processing: PCR fragment-length normalization.
• Paired & non-paired copy-number analysis: All generations, i.e. 10K, 100K, 500K, 5.0 & 6.0. Estimation of full-resolution (raw) copy numbers using CRMA (10K-500K) and CRMAv2 (10K-6.0). Combine data from multiple chip types. Segmentation methods such as CBS, GLAD and HaarSeg.
• Genotyping: CRLMM (100K & 500K)
• Alternative splicing: FIRMA.
• Dynamic HTML reports: ArrayExplorer (HapMap demo), ChromosomeExplorer (HapMap demo, Tumor/Normal Demo). Requires Firefox.
• Design goals: usability, quality & extendibility.