Pathway Fingerprinting
The power of this approach lies in the transformation of a representative set of TSS level activities to a standardized pathway activities.
Data will be posted to this site and updates reminded to the list. Please edit in comments and suggestions.
We will be using fingerprinting in a satellite paper to generate a lineage and functional relationships graph for the FANTOM5 dataset.
Fingerprinting signatures allows us to find related GEO files to the cellular state / network basin. With a fingerprint from a cell state or set of genes we can find the associated GEO array files for any fingerprint (celltype). In this way we can provide broad context for the cell states and potentially, genes involved in network basins.
Fingerprinting provides robust integration across a broad range of experimental platforms, species and data types to identify functional relationships that cannot be discovered by direct TSS, gene or gene signature comparison.
If you wish to - standardize the relationships between cell states - investigate the functions inherent in a cell state or group of cell states - determine relationship between cell types or need a type definition for a cell state - determine which TSS are over-represented in a particular pathway
contact galtschu@hsph.harvard.edu
We are open to collaborations that incorporate our pathway fingerprinting approach.
CAGE TSS profiles generate pathway fingerprints to provide a quantitative functional definition of each cell type. Pathway fingerprints are directly compared across a) the FANTOM5 data corpus b) GEO/Arrayexpress, chemical perturbation libraries, RNAseq and GWAS repositories.
Cell types are classified according to their functional profile. Clustering of sample pathway fingerprints provide a basis for grouping according to shared pathways, within and across species. A simple distance results in a relationship diagram that can be compared with Cell Ontology
Samples will be systematically contextualized within external resources by identifying functionally related experiments in GEO and ArrayExpress, leveraging the vast public data repositories to transfer contextual information. The concept of a sequence search with an unknown gene sequence against an annotated gene database is extended to the concept of a pathway fingerprint search against a known index of fingerprints constructed using public databases.
Lineages will be organized by phylogenetic analysis of a functional distance metric to characterize branch points in differentiation hierarchies. Time-courses will be described in terms of correlated functional transitions.
Pathways will be iteratively expanded by guilt-by-association and the existing data corpus re-analyzed to incorporate this additional functional associations.
