What is Generative Genomics?

The origins of orphan genes have been explored from an evolutionary perspective, with the assumption that they arose either through mutational transformations of genes or as de novo genes derived from non-coding sequences. The current approach of exploring evolutionary origins based on sequence similarity is reasonable for identifying putative origins, but it is not sufficient without simulating the mutational path from the ancestral sequence to the orphan gene including its regulatory elements and functions.
Thus, Generative Genomics research aims to distinguish the origins of orphan genes whether they could have evolved through natural inheritance or were given axiomatically as initial conditions. This endeavor seeks to identify the boundary conditions between evolutionary realm and axiomatic origins.

It extends the traditional scope of comparative genomics, leveraging generative AI to simulate mutational paths, predict phenotypic traits, and build new bioinformatics tools and databases.

Generative Genomics as Comparative Genomics 2.0

Comparative Genomics has traditionally aimed to discover homologous genes and proteins from the perspective of their evolutionary relationships. However, the discovery of orphan genes reveals the potential limitations of purely evolutionary paths. Therefore, it is necessary to distinguish the current homology-based comparative genomics as Comparative Genomics 1.0, in contrast to Comparative Genomics 2.0, which focuses on the study of species-specific and taxon-specific orphan genes and their association with non-heritable phenotypic traits.

Evolutionary and Non-Evolutionary Origins of Orphan Genes

The origins of orphan genes have been explored from an evolutionary perspective, with the assumption that they arose either through mutational transformations of genes or as de novo genes derived from non-coding sequences. The current approach of exploring evolutionary origins based on sequence similarity is reasonable for identifying putative origins, but it is not sufficient without simulating the mutational path from the ancestral sequence to the orphan gene including its regulatory elements and functions.
Thus, Generative Genomics research aims to distinguish the origins of orphan genes whether they could have evolved through natural inheritance or were given axiomatically as initial conditions. This endeavor seeks to identify the boundary conditions between evolutionary realm and axiomatic origins.

Generative Genes in Generative Genomics Study

From a conceptual standpoint, it is necessary to extend the origins of orphan genes beyond evolutionary mechanisms, because it is illogical to assume that all genes evolved from other genes. We must distinguish the genes that existed axiomatically—those from which other genes evolved. Therefore, it is necessary to define the non-evolutionary orphan genes, termed Generative Genes along with their species- and taxon-specific functions. We propose to call this research paradigm Generative Genomics, which utilizes Generative AI, as a branch of Comparative Genomics, and to position it as Comparative Genomics 2.0.

ILLUSTRATIVE TOPICS OF GENERATIVE GENOMICS RESEARCH

To fulfill the objectives of Generative Genomics using Generative AI and to explore its implications for health science, we propose to investigate the following research issues, as illustrated by—but not limited to—the five topics below. They explain the opportunities for researchers studying orphan genes and de novo genes, and how Generative AI models can contribute to the development of Generative Genomics and the identification of species-specific diseases.

Related Topics

1) Origins of Orphan Genes and Generative Genes

• Standard procedures of identifying orphan genes: Sensitivity of e-values and interspecies comparison
• Distinction between evolutionary and non-evolutionary orphan genes (i.e., generative genes)
• Necessary and sufficient conditions for validating the emergence of de novo genes and transformed genes: sequence similarity, regulatory elements, mutational path simulation, and functional integration
• Identification of orphan genes and generative genes across various species

2) Generative Genomics as Comparative Genomics 2.0

• Conceptual distinction between Comparative Genomics 1.0 and 2.0
• Development of automated platforms for identification of generative genes, built upon BLAST, DIAMOND and related tools
• Design of annotation methods for generative genes by mapping them to multi-layered species-specific and taxon-specific phenotypic traits
• Integration of existing annotation databases with a newly annotation database for generative genes

3) Generative Origins of All Organisms in the Generative Tree of Life

• Association of generative genes and unique phenotypic traits in the phylogenetic Tree of Life, leading to the construction of the Generative Tree of Life
• Clarification of the ambiguous concept of ancestors by distinguishing taxa from reproducible organisms
• Identification of the boundary between inherited evolutionary origins and generative origins
• Identification of the network structure of inherited origins in the Generative Tree of Life

4) Generative AI for Generative and Synthetic Genomics

• LLM-based generative AI for the association of multi-omics databases
• LLM-based GenAI models that simulate the new proteins and possibility/impossibility of orphan genes
• Performance comparison of encoder and decoder models in associating generative genes with species-specific phenotypes
• Understanding body growth scheduling under the decentralized cell division mechanism

5) Generative Genomics in Disease Pathway Analysis

• Generative genomics AI model for molecular pathway analysis of diseases
• Discovery of generative gene-specific diseases in humans and other species
• Distinction between treatment methods for diseases originating from common genes and those from generative genes
• Pathological studies using generative genes of C. elegans, fruit flies, and other organisms

Program Committee

• Chair: Prof. Jae Kyu Lee (Xi’an Jiaotong University / KAIST)
• Prof. Dae Kyun Chung (Kyung Hee University)
• Prof. Ah-Ram Kim (Handong Global University)
• Prof. Wooju Kim (Yonsei University)
• Prof. Kyong-Tai Kim (POSTECH)
• Prof. Gyoo Gun Lim (Hanyang University)
• Prof. Taesung Park (Seoul National University)
• Prof. Chuck Yoo (Korea University)
• Prof. Yungang Xu (Xi’an Jiaotong University)