Deep Symbolic Learning Architecture for Variant Calling in NSG

Abstract

[EN]In the era of genomics, efficient and accurate analysis of genomic sequences is essential. Next-generation sequencing (NGS) technology has revolutionised the field of genomics by providing a massive volume of data on an unprecedented scale. One of the critical steps in the analysis of this data is variant calling, where genetic variations are identified from DNA sequences. In this context, we have explored the use of Deep Symbolic Learning (DSL) as an innovative computational approach that combines deep learning with symbolic representations. In this article, we discuss the principles of DSL and its applicability in genomics. We examine the advantages and challenges of its use in the context of variant calling and highlight the importance of meticulous validation. To ensure the quality of the results, it is essential to adopt appropriate validation techniques and specific software tools. We provide a detailed overview of these techniques and tools, with the aim of establishing clear standards for the implementation and validation of DSL algorithms in genomic pipelines. This research highlights the potential of the DSL to improve the accuracy of variant discovery, offering promising prospects for the genomics of the future.