Advances in Artificial Intelligence (AI) Models and Generative Algorithms Represent a New Paradigm for Genomics...

Area of Science:

• Genomics
• Computational Biology
• Bioinformatics

Background:

• Genomic analysis has advanced alongside computing power, moving from basic statistics to sophisticated artificial intelligence (AI).
• Decreasing sequencing costs have generated large, diverse genomic datasets, fueling AI development in the field.
• AI is increasingly applied to predict genomic features like transcription factor binding and DNA methylation.

Purpose of the Study:

• To review the impact of computational progress on genomic analysis.
• To highlight the expanding applications of AI in genomics.
• To discuss the potential of generative AI in simulating and designing genomic sequences.

Main Methods:

• Utilizing advanced AI architectures like deep neural networks, convolutional neural networks, recurrent neural networks, and transformers.
• Integrating multi-omics data to improve predictive accuracy.
• Developing generative AI models for genomic sequence simulation and design.

Main Results:

• AI models now effectively predict transcription factor binding sites, epigenetic elements, DNA methylation, and noncoding sequence functions.
• Modern AI architectures can process longer genomic inputs, capture long-range genomic context, and integrate diverse multi-omics data.
• Generative AI has demonstrated the capability to simulate and design functional genomic sequences.

Conclusions:

• AI, particularly deep learning and generative models, is transforming genomic analysis and interpretation.
• These computational advancements accelerate the understanding and manipulation of genomic data.
• The integration of AI promises to significantly advance our ability to navigate and utilize the genomic landscape.