Skip Navigation

Short Courses

Interface 2007 offers two short courses on Wednesday, May 23, 2007

Computational Systems Biology of Cancer: Measuring, Mining, and Modelling
presented by Bud Mishra (New York University)
Time: 8:00 am to 12:00 noon.
Location: .

Rapid and accurate solutions to many biomedical problems are beginning to rely on systems and computational approaches. A few notable examples are: genomic assays for cancer, genetic analysis of cancer genomes for marker detection, models of cancer progression, etc. While these examples focus on cancer and currently build upon microarray technology, the algorithmic approaches must aim to be scalable, agnostic to the technologies, and applicable to a wide variety of problems. This short course surveys many promises, challenges and obstacles faced by the emerging field of systems biology as they tackle these biomedical problems. In particular, we will emphasize three highly-intertwined aspects of this problem:

  • Measuring: array and single-molecule measurements and their statistical analysis
  • Mining: statistically combining gene expression and genomic patient data for discovery
  • Modeling: systems biology algorithms for reasoning and redescription of time-course data.

Various novel applications of mathematical ideas appear: 0-1-laws in experimental design, scan statistics applications to genetics, rate distortion theory and its applications to model building, hidden Kripke models and temporal logic redescriptions, Nonlinear Kalman-Bucy filtering, efficient maximum a posteriori estimators, etc.

Copy-Number Variation in the Genome: Technologies, Statistics, and Applications
presented by John Marioni, University of Cambridge, Simon Tavare, University of Cambridge and University of Southern California
Time: 1:00 pm to 5:00 pm.
Location: .

Array Comparative Genomic Hybridisation (aCGH) has become an established and important experimental tool for exploring how genomic variation impacts upon many areas of biology. The development of experimental and statistical techniques for analysing aCGH data has been primarily motivated by its application to DNA taken from patients with a particular tumour. However, more recently aCGH has been used to investigate a different problem, namely the extent of Copy Number Variation (CNV) throughout (putatively normal) human genomes. This application has necessitated the development of novel tools for both pre-processing and downstream analysis of the data.

OneĀ  situation where the development of novel tools has been necessary is in the identification of regions of CNV. For tumours, a large number of genomic rearrangements (frequently covering many megabases) often occur and single outlying probes are generally assumed not to represent real biological changes. Moreover, tumours are very heterogeneous and this manifests itself in the presence of many different changes at a particular genomic locus across a cohort of samples. However, when finding regions of CNV using aCGH the problems are quite different. In particular, regions of CNV tend to be small, with an average size of about 200kb. Furthermore, because there are no cellularity effects and sample heterogeneity is less of an issue, it is possible to use cross-sample information to not only remove technical probe effects but also to identify regions of CNV.

In this session, we will describe the development of aCGH as an experimental platform, focussing on the statistical methods that have been devised for analysing the data produced as the technology has evolved. In particular, we will describe the novel methods being used to find regions of CNV and on approaches for analysing regions of CNV in conjunction with gene expression, SNP, methylation and other data generated from the same individuals whose CNV profiles have been determined

Sign up for these courses when you register. See the Registration page for fees.