Omics Data Preprocessing for Machine Learning: A Case Study in Childhood Obesity

Abstract

The use of machine learning techniques for the construction of predictive models of disease outcomes (based on omics and other types of molecular data) has gained enormous relevance in the last few years in the biomedical field. Nonetheless, the virtuosity of omics studies and machine learning tools are subject to the proper application of algorithms as well as the appropriate pre-processing and management of input omics and molecular data. Currently, many of the available approaches that use machine learning on omics data for predictive purposes make mistakes in several of the following key steps: experimental design, feature selection, data pre-processing, and algorithm selection. For this reason, we propose the current work as a guideline on how to confront the main challenges inherent to multi-omics human data. As such, a series of best practices and recommendations are also presented for each of the steps defined. In particular, the main particularities of each omics data layer, the most suitable preprocessing approaches for each source, and a compilation of best practices and tips for the study of disease development prediction using machine learning are described. Using examples of real data, we show how to address the key problems mentioned in multi-omics research (e.g., biological heterogeneity, technical noise, high dimensionality, presence of missing values, and class imbalance). Finally, we define the proposals for model improvement based on the results found, which serve as the bases for future work.

The use of machine learning in paediatric nutrition

“…As with much of the current research into the microbiome, the study focused on developing methods to understand complex interactions between different bacteria. Similarly, a Spanish collaboration used genomic and biochemical predictors of insulin resistance in adolescents to act as a case study [7 ▪ ]. Their study highlighted the importance of careful model selection and data preparation to ensure that reliable and clinically important results were obtained.…”

Integrative multi-omics and systems bioinformatics in translational neuroscience: A data mining perspective

“…Quality control should be applied to all omics layers such as through quantifying GC content in RNA-seq and removing duplicated and fragmented reads for sequence alignment to be performed [ 186 ]. Feature selection is commonly conducted to reduce the dimensionality and redundancy of the high-throughput data, and discriminate desired features contained within the data [ 187 , 188 ]. The goal of the data processing is to reduce dimensionality, bias, and variation of the mined data in order to ensure robustness and efficiency of analysis, especially prior to multi-omics integration.…”

Special Issue: New Advances in Bioinformatics and Biomedical Engineering Using Machine Learning Techniques, IWBBIO-2022

“…In the contribution “Omics Data Preprocessing for Machine Learning: A Case Study in Childhood Obesity”, by Álvaro Torres-Martos et al [ 5 ], the main goal is the analysis of the childhood obesity, which is a multifactorial disease influencing the development of a range of metabolic disorders, where adipose tissue has been proved to be fundamental. The authors present machine learning predictive models with multi-omics human data, and provides a collection of best practices and guidelines that could be applied to other human diseases with complex fundamentals (e.g., obesity).…”