Extraction and Integration of Genetic Networks from Short-Profile Omic Data Sets

Area of Science:

• Systems Biology
• Bioinformatics
• Genomics

Background:

• Mass spectrometry is crucial for profiling intracellular molecules in omics studies.
• Cellular biochemistry and biology are reflected in the covariance of non-Gaussian omics data.
• Understanding correlations in genome-wide mutant libraries is key.

Purpose of the Study:

• Introduce novel similarity measures: Mahalanobis cosine and hybrid Mahalanobis cosine.
• Quantify similarities between profiled features in different mutants using omics data.
• Infer and integrate genetic networks from ionomics/metabolomics data.

Main Methods:

• Utilized Mahalanobis cosine and hybrid Mahalanobis cosine similarity measures.
• Analyzed high-throughput screening omics data, including covariance matrix information.
• Applied measures to ionomics and metabolomics data from *Saccharomyces cerevisiae*.

Main Results:

• The proposed measures effectively quantify similarities between mutant profiles.
• Successfully inferred and integrated multilayer genetic networks for *S. cerevisiae*.
• Demonstrated alternative descriptions of biological relations compared to Pearson's correlation.

Conclusions:

• Mahalanobis cosine measures offer a valuable alternative to traditional correlation coefficients.
• These measures can guide hypothesis generation for uncharacterized genes.
• The study advances the analysis of complex biological systems through omics data integration.