MKL-GRNI: A parallel multiple kernel learning approach for supervised inference of large-scale gene regulatory...

Area of Science:

• Computational biology
• Genomics
• Bioinformatics

Background:

• High-throughput multi-omics data and genomic data fusion enable advanced computational inference models.
• Scalable models can handle large genomes and integrate biological knowledge.
• Sequential execution of these models incurs significant computational costs for large datasets.

Purpose of the Study:

• To present a multiple kernel learning (MKL) based approach for gene regulatory network (GRN) inference.
• To overcome computational limitations in learning from large, heterogeneous genomic datasets.
• To develop a parallel execution architecture for efficient GRN learning.

Main Methods:

• Formulated GRN learning as a supervised classification problem.
• Fused multiple heterogeneous datasets using the MKL paradigm.
• Devised a parallel execution architecture by decomposing the classification problem into subproblems for multi-processor execution.

Main Results:

• The proposed MKL approach demonstrated enhanced speedup and improved inference potential.
• Achieved better classification accuracy compared to state-of-the-art methods.
• Successfully learned large-scale GRNs from multiple, heterogeneous datasets across different species (E. coli, S. cerevisiae, H. sapiens).

Conclusions:

• The MKL-based GRN inference method offers a computationally efficient and accurate solution for large-scale network learning.
• Parallel architecture significantly accelerates the GRN inference process.
• The approach is effective for diverse genomic datasets, advancing computational inference in genomics.