From The MarthLab
[edit] Cancer Genome Analysis Tools
Detection software: Given a detection platform, I have experience in writing data analysis tools that can extract relevant information from data produced by that platform. For example, my current grant determines polymorphic locations in the human genome or heterozygous positions within a single individual on the basis of sequencing data. In the future, I intend to look at other data types e.g. methylation chip data. In that case, we would be writing software that can accurately determine methylation levels along the assayed genome region and provide confidence measures for the accuracy of that determination.
Data integration: Disparate sources of data (e.g. SNPs, somatic mutations, INDELs, copy number changes, chromosomal abnormalities from FISH, expression changes between normal and cancer tissue, methylation changes between normal and cancer tissue) can provide independent evidence to implicate genes as relevant to cancer development. Currently, there are no tools to view and consider these data in a single graphical viewer. Also, when evidence from different data sources are integrated multiple lines of evidence may point to a single entity (gene). Statistically rigorous evaluation of these data is challenging. Another problem is to integrate data from multiple individual cancers of a given type (or of different cancer types, as the case may be). I plan to develop a viewer/statistical analysis tool that can help with these analysesto ultimately pinpoint the genes that are important to go after
