Conservation of Hot Spots and Ligand Binding Sites in Protein Models by AlphaFold2.

The neural network-based program AlphaFold2 (AF2) provides high accuracy structure prediction for a large fraction of globular proteins. An important question is whether these models are accurate enough for reliably docking small ligands. Several recent papers and the results of CASP15 reveal that local conformational errors reduce the success rates of direct ligand docking. Here, we focus on the ability of the models to conserve the location of binding hot spots, regions on the protein surface that significantly contribute to the binding free energy of the protein-ligand interaction. Clusters of hot spots predict the location and even the druggability of binding sites, and hence are important for computational drug discovery. The hot spots are determined by protein mapping that is based on the distribution of small fragment-sized probes on the protein surface and is less sensitive to local conformation than docking. Mapping models taken from the AlphaFold Protein Structure Database show that identifying binding sites is more reliable than docking, but the success rates are still 5% to 10% lower than based on mapping X-ray structures. The drop in accuracy is particularly large for models of multidomain proteins. However, both the model binding sites and the mapping results can be substantially improved by generating AF2 models for the ligand binding domains of interest rather than the entire proteins and even more if using forced sampling with multiple initial seeds. The mapping of such models tends to reach the accuracy of results obtained by mapping the X-ray structures.

Automating biomedical literature review for rapid drug discovery: Leveraging GPT-4 to expedite pandemic response.

OBJECTIVE: The rapid expansion of the biomedical literature challenges traditional review methods, especially during outbreaks of emerging infectious diseases when quick action is critical. Our study aims to explore the potential of ChatGPT to automate the biomedical literature review for rapid drug discovery. MATERIALS AND METHODS: We introduce a novel automated pipeline helping to identify drugs for a given virus in response to a potential future global health threat. Our approach can be used to select PubMed articles identifying a drug target for the given virus. We tested our approach on two known pathogens: SARS-CoV-2, where the literature is vast, and Nipah, where the literature is sparse. Specifically, a panel of three experts reviewed a set of PubMed articles and labeled them as either describing a drug target for the given virus or not. The same task was given to the automated pipeline and its performance was based on whether it labeled the articles similarly to the human experts. We applied a number of prompt engineering techniques to improve the performance of ChatGPT. RESULTS: Our best configuration used GPT-4 by OpenAI and achieved an out-of-sample validation performance with accuracy/F1-score/sensitivity/specificity of 92.87%/88.43%/83.38%/97.82% for SARS-CoV-2 and 87.40%/73.90%/74.72%/91.36% for Nipah. CONCLUSION: These results highlight the utility of ChatGPT in drug discovery and development and reveal their potential to enable rapid drug target identification during a pandemic-level health emergency.