Introduction
In the field of bioinformatics, the interpretation of large gene lists derived from high-throughput genomic or proteomic studies is a challenging task. Many publicly available bioinformatics tools are used for gene list functional enrichment analysis and functional annotation, including but not limited to GoMiner, DAVID, EASE, and GOstat.DAVID is a popular bioinformatics resource system that provides a module-centered approach for functional annotation and enrichment analysis of gene lists. The tool is designed to compress and organize gene lists into biologically meaningful categories, called biomodules, to facilitate the interpretation of gene data in a networked environment.
Fig. 1. DAVID: a web server for functional enrichment analysis and functional annotation of gene lists (2021 update). (Sherman B T, et al., 2022)
Rationale
Individual genes are obviously associated with multiple biological terms. These associations form a complex network of "multi-gene-to-multi-term" relationships that represent the truly complex nature of biological processes. A data mining tool capable of extracting these complex and redundant relationships should be able to recognize functional gene term biological modules.
The DAVID Gene Function Classification tool (http://david.abcc.ncifcrf.gov) employs a novel aggregation algorithm to mine complex biological co-occurrences found in multiple functional annotation sources. By utilizing this algorithm, the tool can group functionally relevant genes and terms into a manageable number of biomodules, thus enabling researchers to efficiently interpret their gene lists.
Main Process of the DAVID Gene Function Classification Tool
The DAVID Gene Functional Classification Tool consists of three main steps: measuring the functional relationships of gene pairs, classifying genes into functional genomes using a clustering method, and visualizing the results in text and graphical modes.
Measuring Functional Relationships Between Gene Pairs Based on Global Annotation Mapping
To measure the functional relationships between gene pairs, the tool compiles a gene term annotation matrix in binary mode containing thousands of annotated terms from a variety of sources. The matrix indicates the presence or absence of a particular annotation term for each gene. The Kappa statistic is a co-occurrence chance-corrected measure of annotation co-occurrence between gene pairs. The measure takes into account the binary nature of annotation profiles.
Annotation terms are organized into flat linear sets to account for redundancy and structured relationships. This approach simplifies the situation and maximizes the use of heterogeneous annotations in the similarity measure. Functional similarity measurements based on annotation profiles of gene pairs can identify genes that share major biological features, even if they do not share significant sequence similarity.
The DAVID Clustering Method Classifies Genes into Functional Gene Groups
The aggregation method in the DAVID Gene Functional Classification Tool groups related genes or terms into functional groups, called biological modules, based on measured similarity distances. Unlike exclusionary methods such as hierarchical, K-means, or SOM clustering, aggregation methods allow a gene or term to participate in multiple functional groups, reflecting the true nature of genes with multiple roles.
Functional groups are ranked according to the overall participation of group members in the enriched biological processes associated with the gene list. The ranking provides insights into the most relevant biological processes represented by the gene list. In addition, a unique fuzzy heatmap visualization provides a global view of the relationships between different groups.
Visualization of Results in Text and Graphical Modes
The utilization of the DAVID Gene Function Classification tool empowers researchers with a diverse range of visualization options, augmenting their ability to explore and comprehend the obtained results. This sophisticated tool encompasses a subset of the acclaimed "drill-down" functionality, enabling investigators to delve deeper into the intricate relationships that exist between genes and terms within the biological modules. By harnessing this feature, a more intricate and exhaustive analysis of functional connections and associations becomes attainable.
Moreover, the tool encompasses both text and graphical modes, affording researchers the flexibility to access and comprehend information through distinct avenues. The text mode presents results in a concise and meticulously organized manner, facilitating efficient assimilation of critical findings. Conversely, the graphical mode employs various visual representations, such as heat maps, to elucidate gene-to-gene, term-to-term, and gene-to-term relationships, thereby aiding researchers in extracting meaningful insights from complex data structures.
Conclusion
In conclusion, the DAVID Gene Function Classification tool stands as an indispensable resource for conducting functional analysis and interpreting vast gene lists. Leveraging its unique aggregation algorithms and employing intricate biological co-occurrence mining techniques, this tool effectively organizes extensive gene lists into biologically coherent modules. It empowers researchers to comprehensively explore intricate gene-term relationships, identify and comprehend enriched biological processes, and gain profound understanding into the functional significance of their gene lists.
Addressing the inherent challenges associated with the interpretation of large gene lists, the DAVID Gene Function Classification tool serves as a user-friendly web server that embraces a module-centric approach. Its formidable algorithms, coupled with its powerful visualization capabilities and comprehensive functionality, position it as an invaluable asset for researchers operating in the ever-evolving field of bioinformatics.
References
- Sherman B T, Hao M, Qiu J, et al. DAVID: a web server for functional enrichment analysis and functional annotation of gene lists (2021 update)[J]. Nucleic acids research, 2022, 50(W1): W216-W221.
- Huang D W, Sherman B T, Tan Q, et al. The DAVID Gene Functional Classification Tool: a novel biological module-centric algorithm to functionally analyze large gene lists[J]. Genome biology, 2007, 8(9): 1-16.
