Matching experiments across species using expression values and textual information

Learning a genome-wide score of human-mouse conservation at the functional genomics level

Identifying genomic regions with functional genomic properties that are conserved between human and mouse is an important challenge in the context of mouse model studies. To address this, we develop a method to learn a score of evidence of conservation at the functional genomics level by integrating information from a compendium of epigenomic, transcription factor binding, and transcriptomic data from human and mouse. The method, Learning Evidence of Conservation from Integrated Functional genomic annotations (LECIF), trains neural networks to generate this score for the human and mouse genomes. The resulting LECIF score highlights human and mouse regions with shared functional genomic properties and captures correspondence of biologically similar human and mouse annotations. Analysis with independent datasets shows the score also highlights loci associated with similar phenotypes in both species. LECIF will be a resource for mouse model studies by identifying loci whose functional genomic properties are likely conserved.

Comparison of Gene Coexpression Profiles and Construction of Conserved Gene Networks to Find Functional Modules

Background

Computational approaches toward gene annotation are a formidable challenge, now that many genome sequences have been determined. Each gene has its own function, but complicated cellular functions are achieved by sets of genes. Therefore, sets of genes with strong functional relationships must be identified. For this purpose, the similarities of gene expression patterns and gene sequences have been separately utilized, although the combined information will provide a better solution.

Result & Discussion

We propose a new method to find functional modules, by comparing gene coexpression profiles among species. A coexpression pattern is represented as a list of coexpressed genes with each guide gene. We compared two coexpression lists, one from a human guide gene and the other from a homologous mouse gene, and defined a measure to evaluate the similarity between the lists. Based on this coexpression similarity, we detected the highly conserved genes, and constructed human gene networks with conserved coexpression between human and mouse. Some of the tightly coupled genes (modules) showed clear functional enrichment, such as immune system and cell cycle, indicating that our method could identify functionally related genes without any prior knowledge. We also found a few functional modules without any annotations, which may be good candidates for novel functional modules. All of the comparisons are available at the http://v1.coxsimdb.info web database.

Toward computational cumulative biology by combining models of biological datasets

A main challenge of data-driven sciences is how to make maximal use of the progressively expanding databases of experimental datasets in order to keep research cumulative. We introduce the idea of a modeling-based dataset retrieval engine designed for relating a researcher's experimental dataset to earlier work in the field. The search is (i) data-driven to enable new findings, going beyond the state of the art of keyword searches in annotations, (ii) modeling-driven, to include both biological knowledge and insights learned from data, and (iii) scalable, as it is accomplished without building one unified grand model of all data. Assuming each dataset has been modeled beforehand, by the researchers or automatically by database managers, we apply a rapidly computable and optimizable combination model to decompose a new dataset into contributions from earlier relevant models. By using the data-driven decomposition, we identify a network of interrelated datasets from a large annotated human gene expression atlas. While tissue type and disease were major driving forces for determining relevant datasets, the found relationships were richer, and the model-based search was more accurate than the keyword search; moreover, it recovered biologically meaningful relationships that are not straightforwardly visible from annotations—for instance, between cells in different developmental stages such as thymocytes and T-cells. Data-driven links and citations matched to a large extent; the data-driven links even uncovered corrections to the publication data, as two of the most linked datasets were not highly cited and turned out to have wrong publication entries in the database.