Are we far from correctly inferring gene interaction networks with Lasso?

TL;DR

This paper reviews nine penalised regression methods for inferring gene interaction networks from microarray data, highlighting their limitations and suggesting precautions to improve prediction reliability.

Contribution

It provides a comprehensive evaluation of existing penalised regression methods for gene network inference and discusses their limitations and best practices.

Findings

Different methods vary in handling biological data complexity.

Limitations of each method can affect prediction accuracy.

Recommendations for more reliable gene network inference.

Abstract

Detecting the interactions of genetic compounds like genes, SNPs, proteins, metabolites, etc. can potentially unravel the mechanisms behind complex traits and common genetic disorders. Several methods have been taken into consideration for the analysis of different types of genetic data, regression being one of the most widely adopted. Without any doubt, a common data type is represented by gene expression profiles, from which gene regulatory networks have been inferred with different approaches. In this work we review nine penalised regression methods applied to microarray data to infer the topology of the network of interactions. We evaluate each method with respect to the complexity of biological data. We analyse the limitations of each of them in order to suggest a number of precautions that should be considered to make their predictions more significant and reliable.