‘Bioinformatics for statistical and functional analysis of genome-wide expression data from cancer clinical samples’

Alberto de luis Calvo, Harel Weinstein, Javier De las Rivas Sanz, Jesús López Fidalgo, Norma Gutiérrez Gutiérrez.

Human genomic studies are being carried out in the Cancer Research Center (CIC, Salamanca) using high-density oligo-nucleotide microarrays with samples obtained from patients with different types of cancer. These microarrays provide gene expression data from the complete human genome for each analyzed sample. These experimental studies are producing a massive amount of data which have to be processed with robust and systematic methods to yield results with full statistical and biological meaning.

Using bioinformatic methods we will evaluate and develop a series of steps needed to achieve thorough data analyses and to extract the most biological information that these genomic data contain.The steps are: 1) robust normalization using intra-chip and inter-chip comparison; 2) implementation of algorithms to select the genes which present a statistically significant differential expression; 3) evaluation of clustering methods to select the samples with similar expression patterns and to identify groups within the patients (class discovery); 4) identification of the genes that best define the expression profile or pattern of each class (class prediction); 5) development and application of bioinformatic strategies to identify gene-networks from genome-wide expression data; 6) identification of some specific functional networks that may involve some key genes, which can be special targets for cancer research.