Generative artificial intelligence performs rudimentary structural biology modeling.

Natural language-based generative artificial intelligence (AI) has become increasingly prevalent in scientific research. Intriguingly, capabilities of generative pre-trained transformer (GPT) language models beyond the scope of natural language tasks have recently been identified. Here we explored how GPT-4 might be able to perform rudimentary structural biology modeling. We prompted GPT-4 to model 3D structures for the 20 standard amino acids and an α-helical polypeptide chain, with the latter incorporating Wolfram mathematical computation. We also used GPT-4 to perform structural interaction analysis between nirmatrelvir and its target, the SARS-CoV-2 main protease. Geometric parameters of the generated structures typically approximated close to experimental references. However, modeling was sporadically error-prone and molecular complexity was not well tolerated. Interaction analysis further revealed the ability of GPT-4 to identify specific amino acid residues involved in ligand binding along with corresponding bond distances. Despite current limitations, we show the capacity of natural language generative AI to perform basic structural biology modeling and interaction analysis with atomic-scale accuracy.

Towards multi-target glioblastoma therapy: Structural, distribution, and functional insights into protein target candidates.

Glioblastoma is the most commonly occurring and most lethal primary brain tumor. Treatment options are limited in number and therapeutic development remains a major challenge. However, substantial progress has been made in better understanding the underlying biology of the disease. A recent proteomic meta-analysis revealed that 270 proteins were commonly dysregulated in glioblastoma, highlighting the complexity of the disease. This motivated us to explore potential protein targets which may be collectively inhibited, based on common upregulation, as part of a multi-target therapeutic strategy. Herein, we identify and characterize structural attributes relevant to the druggability of six protein target candidates. Computational analysis of crystal structures revealed druggable cavities in each of these proteins, and various parameters of these cavities were determined. For proteins with inhibitor-bound structures available, inhibitor compounds were found to overlap with the computationally determined cavities upon structural alignment. We also performed bioinformatic analysis for normal transcriptional expression distribution of these proteins across various brain regions and various tissues, as well as gene ontology curation to gain functional insights, as this information is useful for understanding the potential for off-target adverse effects. Our findings represent initial steps towards the development of multi-target glioblastoma therapy and may aid future work exploring similar therapeutic strategies.

AI interprets the Central Dogma and Genetic Code.

Generative artificial intelligence (AI) is a burgeoning field with widespread applications, including in science. Here, we explore two paradigms that provide insight into the capabilities and limitations of Chat Generative Pre-trained Transformer (ChatGPT): its ability to (i) define a core biological concept (the Central Dogma of molecular biology); and (ii) interpret the genetic code.

The Central Dogma revisited: Insights from protein synthesis, CRISPR, and beyond.

Francis Crick advanced two distinct but interrelated fundamental principles of molecular biology: (1) the Sequence Hypothesis and (2) the Central Dogma. The Sequence Hypothesis defines biological information transfer as the residue-by-residue transfer of sequence information between nucleic acids and to proteins. This is commonly summarized as DNA âž” RNA âž” protein and is colloquially referred to as the Central Dogma. More specifically, however, the Central Dogma expounded by Crick included a critical restriction, stipulating that "once sequential information has passed into protein it cannot get out again." Under this definition, the Central Dogma has stood the test of time despite challenges. In principle, a violation of the Central Dogma could transpire through synthetic biology or by natural occurrence. To address these possibilities, we draw insights from existing modes of information transfer in protein synthesis and from synthetic Clustered Regularly-Interspaced Short Palindromic Repeats (CRISPR) gene-editing. We introduce a three-part evaluation scheme, which we apply to the CRISPR/Cas9 system and the more recent CRISPR prime editing system. Potential mechanisms by which engineered sequence editing systems might violate the Central Dogma are considered. We conclude that although information transfer in protein synthesis and CRISPR gene-editing remain within the bounds of the Central Dogma, the underlying mechanisms point toward an avenue of synthetic biology that could directly violate the Central Dogma. Finally, we speculate on some of the theoretical and practical implications of a protein-derived information transfer system. This article is categorized under: RNA Evolution and Genomics > Ribonomics RNA Interactions with Proteins and Other Molecules > Protein-RNA Interactions: Functional Implications Translation > Mechanisms.

Protein LY6E as a candidate for mediating transport of adeno-associated virus across the human blood-brain barrier.

The blood-brain barrier (BBB) is a major obstacle for the treatment of central nervous system (CNS) disorders. Significant progress has been made in developing adeno-associated virus (AAV) variants with increased ability to cross the BBB in mice. However, these variants are not efficacious in non-human primates. Herein, we employed various bioinformatic techniques to identify lymphocyte antigen-6E (LY6E) as a candidate for mediating transport of AAV across the human BBB based on the previously determined mechanism of transport in mice. Our results provide insight into future discovery and optimization of AAV variants for CNS gene delivery in humans.