Andrea Townsend-Nicholson

Therapeutic interventions for complex diseases depend on the targeted modulation of key pathological pathways. While growing clinical needs continue to drive advancements in the drug discovery space, current strategies primarily rely on searching large volumes of chemical data without addressing the specific contributions of molecular features. Moreover, both clinicians and researchers recognize the need for improved drug discovery methods and characterization that could aid in clinical strategy selection. To address these challenges, we propose a new perspective on targeted therapy development as well as interactome mapping, utilizing molecular fragments. The present study focuses on therapeutic areas that represent emerging targets, namely JAK2 and GLP-1R, both of which have broad clinical potential. We developed a new self-adjusting neural network that enabled us to discover novel therapeutic candidates with improved in silico binding profiles, gain additional insights into drug-target binding that were not previously reported, and identify new metabolic trajectories. Importantly, our work revealed that even a small compound library can effectively generate lead candidates, expediting the search and exploration process. In addition, the fragment-guided bridging of chemical and biological spaces has revealed new opportunities for drug repurposing efforts and a means of improving the prediction of side effects. We concluded our study with insights into the recent high-profile clinical trial failure of danuglipron and how this could have been prevented with our methodology. Thus, building a robust in silico pipeline with integrated screening data can significantly reduce costs and guide therapy adoption. Furthermore, our proposed strategy highlights promising avenues for the discovery of new therapeutics and the development of clinical interventions.

Therapeutic interventions for complex diseases depend on the targeted modulation of key pathological pathways. While growing clinical needs continue to drive advancements in the drug discovery space, current strategies primarily rely on searching large volumes of chemical data without addressing the specific contributions of molecular features. Moreover, both clinicians and researchers recognize the need for improved drug discovery methods and characterization that could aid in clinical strategy selection. To address these challenges, we propose a new perspective on targeted therapy development as well as interactome mapping, utilizing molecular fragments. The present study focuses on therapeutic areas that represent emerging targets, namely JAK2 and GLP-1R, both of which have broad clinical potential. We developed a new self-adjusting neural network that enabled us to discover novel therapeutic candidates with improved in silico binding profiles, gain additional insights into drug-target binding that were not previously reported, and identify new metabolic trajectories. Importantly, our work revealed that even a small compound library can effectively generate lead candidates, expediting the search and exploration process. In addition, the fragment-guided bridging of chemical and biological spaces has revealed new opportunities for drug repurposing efforts and a means of improving the prediction of side effects. We concluded our study with insights into the recent high-profile clinical trial failure of danuglipron and how this could have been prevented with our methodology. Thus, building a robust in silico pipeline with integrated screening data can significantly reduce costs and guide therapy adoption. Furthermore, our proposed strategy highlights promising avenues for the discovery of new therapeutics and the development of clinical interventions. TOC Graphic:

Structure-based drug design (SBDD) is rapidly evolving to be a fundamental tool for faster and more cost-effective methods of lead drug discovery. SBDD aims to offer a computational replacement to traditional high-throughput screening (HTS) methods of drug discovery. This “virtual screening” technique utilizes the structural data of a target protein in conjunction with large databases of potential drug candidates and then applies a range of different computational techniques to determine which potential candidates are likely to bind with high affinity and efficacy. It is proposed that high-throughput SBDD (HT-SBDD) will significantly enrich the success rate of HTS methods, which currently fluctuates around ~1%. In this chapter, we focus on the theory and utility of high-throughput drug docking, fragment molecular orbital calculations, and molecular dynamics techniques. We also offer a comparative review of the benefits and limitations of traditional methods against more recent SBDD advances. As HT-SBDD is computationally intensive, we will also cover the important role high-performance computing (HPC) clusters play in the future of computational drug discovery. Key words Structure-based drug design Drug development Ligand docking Fragment molecular orbitals FMO Molecular dynamics High-performance computing Virtual screening

At the “Kick Off” meeting for CompBioMed (compbiomed.eu), which was first funded in October 2016, I had no idea that one single sentence (“I wish I could teach this to medical students”) would lead to a dedicated program of work to engage the clinicians and biomedical researchers of the future with supercomputing. This program of work which, within the CompBiomed Centre of Excellence, we have been calling “the CompBioMed Education and Training Programme,” is a holistic endeavor that has been developed by and continues to be delivered with the expertise and support from experimental researchers, computer scientists, clinicians, HPC centers, and industrial partners within or associated with CompBioMed. The original description of the initial educational approach to training has previously been published (Townsend-Nicholson Interface Focus 10:20200003, 2020). In this chapter, I describe the refinements to the program and its delivery, emphasizing the highs and lows of delivering this program over the past 6 years. I conclude with suggestions for feasible measures that I believe will help overcome the barriers and challenges we have encountered in bringing a community of users with little familiarity of computing beyond the desktop to the petascale and beyond. Key words High-performance computing University education Medical student Undergraduate Molecular biosciences Experimental-computational workflow Metagenomics Next-generation sequencing Computational biology Computational biomedicine

The G protein-coupled receptor Smoothened (Smo) is a central signal transducer of the Hedgehog (Hh) pathway which has been linked to diverse forms of tumours. Stimulated by advancements in structural and functional characterisation, the Smo receptor has been recognised as an important therapeutic target in Hh-driven cancers, and several Smo inhibitors have now been approved for cancer therapy. This receptor is also known to be an oncoprotein itself and its gain-of-function variants have been associated with skin, brain, and liver cancers. According to the COSMIC database, oncogenic mutations of Smo have been identified in various other tumours, although their oncogenic effect remains unknown in these tissues. Drug resistance is a common challenge in cancer therapies targeting Smo, and data analysis shows that healthy individuals also harbour resistance mutations. Based on the importance of Smo in cancer progression and the high incidence of resistance towards Smo inhibitors, this review suggests that detection of Smo variants through tumour profiling could lead to increased precision and improved outcomes of anti-cancer treatments.

The G protein-coupled receptor (GPCR) Smoothened (Smo) is a central signal transducer of the Hedgehog (Hh) pathway which has been linked to diverse forms of tumours. Stimulated by advancements in structural and functional characterisation, the Smo receptor has been recognised as an important therapeutic target in Hh-driven cancers, and several Smo inhibitors have now been approved for cancer therapy. This receptor is also known to be an oncoprotein itself and its gain-of-function variants have been associated with skin, brain, and liver cancers. According to the COSMIC database, oncogenic mutations of Smo have been identified in various other tumours, although their oncogenic effect remains unknown in these tissues. Drug resistance is a common challenge in cancer therapies targeting Smo, and data analysis shows that healthy individuals also harbour resistance mutations. Based on the importance of Smo in cancer progression and the high incidence of resistance towards Smo inhibitors, this review suggests that detection of Smo variants through tumour profiling could lead to increased precision and improved outcomes of anti-cancer treatments.

The final year of a biochemistry degree is usually a time to experience research. However, laboratory‐based research projects were not possible during COVID‐19. Instead, we used open datasets to provide computational research projects in metagenomics to biochemistry undergraduates (80 students with limited computing experience). We aimed to give the students a chance to explore any dataset, rather than use a small number of artificial datasets (~60 published datasets were used). To achieve this, we utilized Google Colaboratory (Colab), a virtual computing environment. Colab was used as a framework to retrieve raw sequencing data (analyzed with QIIME2) and generate visualizations. Setting up the environment requires no prior experience; all students have the same drive structure and notebooks can be shared (for synchronous sessions). We also used the platform to combine multiple datasets, perform a meta‐analysis, and allowed the students to analyze large datasets with 1000s of subjects and factors. Projects that required increased computational resources were integrated with Google Cloud Compute. In future, all research projects can include some aspects of reanalyzing public data, providing students with data science experience. Colab is also an excellent environment in which to develop data skills in multiple languages (e.g., Perl, Python, Julia).

Protein-protein interactions (PPIs) are essential for the function of many proteins. Aberrant PPIs have the potential to lead to disease, making PPIs promising targets for drug discovery. There are over 64,000 PPIs in the human interactome reference database; however, to date, very few PPI modulators have been approved for clinical use. Further development of PPI-specific therapeutics is highly dependent on the availability of structural data and the existence of reliable computational tools to explore the interface between two interacting proteins. The fragment molecular orbital (FMO) quantum mechanics method offers comprehensive and computationally inexpensive means of identifying the strength (in kcal/mol) and the chemical nature (electrostatic or hydrophobic) of the molecular interactions taking place at the protein-protein interface. We have integrated FMO and PPI exploration (FMO-PPI) to identify the residues that are critical for protein-protein binding (hotspots). To validate this approach, we have applied FMO-PPI to a dataset of protein-protein complexes representing several different protein subfamilies and obtained FMO-PPI results that are in agreement with published mutagenesis data. We observed that critical PPIs can be divided into three major categories: interactions between residues of two proteins (intermolecular), interactions between residues within the same protein (intramolecular), and interactions between residues of two proteins that are mediated by water molecules (water bridges). We extended our findings by demonstrating how this information obtained by FMO-PPI can be utilized to support the structure-based drug design of PPI modulators (SBDD-PPI).

Drug-target residence time, the duration of binding at a given protein target, has been shown in some protein families to be more significant for conferring efficacy than binding affinity. To carry out efficient optimization of residence time in drug discovery, machine learning models that can predict that value need to be developed. One of the main challenges with predicting residence time is the paucity of data. This chapter outlines all of the currently available ligand kinetic data, providing a repository that contains the largest publicly available source of GPCR-ligand kinetic data to date. To help decipher the features of kinetic data that might be beneficial to include in computational models for the prediction of residence time, the experimental evidence for properties that influence residence time are summarized. Finally, two different workflows for predicting residence time with machine learning are outlined. The first is a single-target model trained on ligand features; the second is a multi-target model trained on features generated from molecular dynamics simulations.