生物信息学概论组与生物信息学II.ppt

Cell Location Robust Genes Sensitive Genes Enriched Cellular Compartment P value Enriched Cellular Compartment P value intracellular 1.36e-12 extracellular space 1.11e-18 ribosomal subunit 4.9e-11 extracellular region 4.4e-18 nucleus 1.12e-08 MHC class II protein complex 7.57e-05 Actors in networks Transcriptional regulatory network PPI Network Transcriptional regulatory network More transcription factors were found for the sensitive genes. 26 TFs for robust, 155 for sensitive In protein-interaction networks Limitations The HGU133A just represent part of human genes. The quality of PPI network and Transcriptional regulatory network. Data bias may affect the result. Conclusion expectation Our analysis suggests that genes have different propensities corresponding to perturbations and such propensities should be considered in the gene expression data analysis. Understanding gene expression sensitivity has important implications for choosing biomarkers, drug targets etc from transcriptional profiling data. With more datasets, this algorithm will help to do more concrete analysis. Difficulties in studying transcriptome? - Huge amount of data and complexity associated with them - Multiple platforms: Affy, Agilent, NGS Variation by sample preparation Noise vs signal Complex algorithms: RMA, MAS5, dChip, PLIER, VSN, PDNN Expensive and exclusive analysis tools How can we make it feasible to compare expression profiles of cells, tissues, or organisms cross multiple experiments? How can we make meta-analysis of transcriptomes practical and meaningful? A Gene Module Based Meta-analysis of Gene Expression Data Goals: Develop gene module based abstraction method, in order to enable meta-analysis of gene expression data cross experiments and cross platforms Build reference expression profile database for model organisms Provide a gene module based expression similarity search tool: GEMBlast (which can be used similar to sequnce BLAST search) Instance：