MLPHENOM

Machine learning for quantitative modelling of structured phenotypes

 Partecipanti

Mappa

 Word cloud

Esplora la "nuvola delle parole (Word Cloud) per avere un'idea di massima del progetto.

 Obiettivo del progetto (Objective)

'Understanding phenotypic variation, and more particularly identifying the causal genetic or environmental regulators, is a major aim in biological investigations. The goal of this proposal is to develop and apply machine learning techniques to model key aspects of structure that occur in modern, high-dimensional phenotype datasets. First, the temporal structure of phenotypes that are recorded over time is addressed. Statistical models can exploit smoothness of time series and detect change points. Second, the structure of images, arising when digital pictures are used as phenotypic variables, is considered. Machine learning techniques allow interpretable image features to be automatically extracted and used as quantitative traits, complementing classical measurements. Finally, the network structure of the phenome is addressed. Different phenotype variables influence each other, resulting in a chain of effects that needs to be modelled to reveal the true causal relationships. The developed algorithms will be applied to understand phenotypic variation in Arabidopsis thaliana in direct collaboration with researchers at the Max Planck Institute for Developmental Biology.'

Introduzione (Teaser)

A prime objective of genetic studies is the identification of disease-causing mutations and how they result in disease. European researchers have provided a helping hand by generating multivariate statistical models.

Descrizione progetto (Article)

Genetic differences among individuals are translated into phenotypic differences. However, most disease phenotypes emerge not from single genes and proteins, but from a complex network of molecular interactions. Each phenotype variable often influences the other, resulting in a domino effect that needs to be modelled to reveal the true causal relationships.

To address this, scientists on the EU-funded 'Machine learning for quantitative modelling of structured phenotypes' (MLPHENOM) project proposed to develop statistical and computational methods for interpreting genotype to phenotype relationships. This entailed analysis of repeated measurements of the same phenotype over time, of digital images, and of networks implicated in a particular phenotype.

In this context, the consortium applied machine learning techniques to model time, image and network aspects of phenotype. Machine learning methods extract certain features of images and use them as quantitative phenotypic traits, thereby complementing classical measurements.

The statistical framework generated during MLPHENOM could take into account hundreds of individual measurements for the analysis of phenotype networks. This allowed researchers to address diverse aspects of phenotype structure and also detect the potential implication of environmental factors. Additional models enabled the stratification of individual molecular events that lead from genotype to phenotype.

MLPHENOM's model system laid the foundation for the analysis of future genetic studies involving large numbers of complex phenotypes. Especially in biomedical research, this modelling approach is expected to find wide applicability.