Kanonte, Mariam and Covenant University, Theses (2024) A MULTI-CRITERIA RANKING SYSTEM FOR EVALUATING GENE REGULATORY NETWORK INFERENCE METHODS. Masters thesis, Covenant University.
![[img]](http://eprints.covenantuniversity.edu.ng/style/images/fileicons/application_pdf.png) |
PDF
Download (234kB) |
Abstract
In this study, a multi-criteria ranking algorithm was introduced for evaluating gene regulatory network inference (GRNi) methods called RankGRN. This algorithm integrates a technique for determining the weights of evaluation criteria based on priorities defined by the user for ranking GRNi methods. Built upon the Weighted Sum Model, RankGRN, first generates the weight of each evaluation criteria considering the user-specified priority, after which it calculates the score of each GRNi method from all the selected weighted evaluation criteria, then it ranks the GRNi methods and identifies the best method according to the priority of each evaluation criteria and the overall best method. Subsequently, a web-based framework was developed for benchmarking and ranking GRNi methods called GRN Evaluator. The newly developed multi-criteria ranking algorithm was integrated in GRN Evaluator for ranking the GRNi methods after benchmarking. GRN Evaluator is a novel framework designed to address limitations in existing GRN benchmarking frameworks such as BEELINE, GReNaDIne, NetBenchmark, etc. Key advancements in GRN Evaluator include the incorporation of machine learning methods for GRN inference, addition of more datasets (both single-cell and bulk RNA-Seq data), inclusion of additional evaluation metrics, and integrating visualization tools for better interpretation of gene networks. Additionally, GRN Evaluator offers a user-friendly web interface, enhancing accessibility and usability. The systematic approach to evaluating GRNi methods across multiple datasets used in this study demonstrates their performance in various contexts. The framework effectively ranks these methods, providing valuable insights for researchers. The findings from this study serve as a guideline for selecting appropriate GRNi methods based on users’ specific needs.
| Item Type: | Thesis (Masters) |
|---|---|
| Uncontrolled Keywords: | gene regulatory network inference methods, machine learning algorithms, ranking algorithm, multi-criteria decision making, single cell RNA-Seq, bulk RNA-Seq, benchmarking frameworks |
| Subjects: | Q Science > QA Mathematics Q Science > QA Mathematics > QA75 Electronic computers. Computer science |
| Divisions: | Faculty of Engineering, Science and Mathematics > School of Electronics and Computer Science |
| Depositing User: | nwokealisi |
| Date Deposited: | 09 Apr 2024 13:32 |
| Last Modified: | 09 Apr 2024 13:32 |
| URI: | http://eprints.covenantuniversity.edu.ng/id/eprint/17891 |
Actions (login required)
 |
View Item |
