GWAS studies compare the DNA of people with a given disease against that of people without it, analyzing millions of genetic variants to identify those associated with the disease. The project responded to the need to systematize and speed up this work, which requires processing large volumes of data with high computing capacity.
The project's objectives were:
- To evaluate the performance of existing methodologies for GWAS studies, both statistical and artificial intelligence-based.
- To design, develop, and evaluate a new artificial intelligence-based methodology that would complement and improve on the techniques analyzed.
- To study parallelization techniques applicable to GWAS studies.
AMALGAMA was structured around the following areas:
- Test genomic data: generation of a synthetic genomic dataset to validate the ability of the methodologies to detect the truly significant SNPs.
- Comparative evaluation of methodologies: analysis of univariate and multivariate statistical methods and machine learning algorithms, validated on both synthetic data and a real genomic dataset.
- New methodology and parallelization: design and development of a proprietary artificial intelligence-based method and study of parallelization techniques to reduce computing times.
CTIC carried out the project by applying its capabilities in intelligent data analysis and artificial intelligence to the study of the human genome, with a methodology transferable to any other scenario that requires analyzing large volumes of data with high computing capacity.
