Genomic Sequence Generation and Drug Discovery using NLP and Generative AI: A Review

Modern artificial intelligence developments have revolutionized genomics and drug discoveryresearchbyusingNaturalLanguageProcessingandGenerativeAItechnologies.The primarychallengeinvolvestwotaskswhichrequirescientiststoanalyzeextensivegenomic data and create biologically accurate genetic sequences.The review examines genomic sequence modeling through three different types of models which include transformer-based modelsanddiffusionmodelsandautoregressivearchitectures.Theresearchdemonstrates thatDNAsequencesfunctionaslanguagewhichallowsfortheprocessoftokenizationand the development of contextual understanding and sequence creation.The research results demonstratethatlargelanguagemodels(LLMs)enhancetheabilitytopredictregulatory elements and interpret variants and create synthetic sequences.The drug discovery processusesgenerativemodelstocreatemolecularstructuresandidentifybiologicaltargetsand forecastproteinstructures.Thefieldofpersonalizedmedicineusesbioinformaticsautomationtocreatediseasepredictionsystems.Thereviewidentifiesseverallimitationswhich prevent complete understanding of the research results while putting forward multimodal learning and explainable AI as future research paths.