Dissecting celastrol with machine learning to unveil dark pharmacology

Evaluation guidelines for machine learning tools in the chemical sciences

Machine learning (ML) promises to tackle the grand challenges in chemistry and speed up the generation, improvement and/or ordering of research hypotheses. Despite the overarching applicability of ML workflows, one usually finds diverse evaluation study designs. The current heterogeneity in evaluation techniques and metrics leads to difficulty in (or the impossibility of) comparing and assessing the relevance of new algorithms. Ultimately, this may delay the digitalization of chemistry at scale and confuse method developers, experimentalists, reviewers and journal editors. In this Perspective, we critically discuss a set of method development and evaluation guidelines for different types of ML-based publications, emphasizing supervised learning. We provide a diverse collection of examples from various authors and disciplines in chemistry. While taking into account varying accessibility across research groups, our recommendations focus on reporting completeness and standardizing comparisons between tools. We aim to further contribute to improved ML transparency and credibility by suggesting a checklist of retro-/prospective tests and dissecting their importance. We envisage that the wide adoption and continuous update of best practices will encourage an informed use of ML on real-world problems related to the chemical sciences.

Natural product drug discovery in the artificial intelligence era

Natural products (NPs) are primarily recognized as privileged structures to interact with protein drug targets. Their unique characteristics and structural diversity continue to marvel scientists for developing NP-inspired medicines, even though the pharmaceutical industry has largely given up. High-performance computer hardware, extensive storage, accessible software and affordable online education have democratized the use of artificial intelligence (AI) in many sectors and research areas. The last decades have introduced natural language processing and machine learning algorithms, two subfields of AI, to tackle NP drug discovery challenges and open up opportunities. In this article, we review and discuss the rational applications of AI approaches developed to assist in discovering bioactive NPs and capturing the molecular "patterns" of these privileged structures for combinatorial design or target selectivity.

Artificial intelligence-guided discovery of anticancer lead compounds from plants and associated microorganisms

Plants and associated microorganisms are essential sources of natural products against human cancer diseases, partly exemplified by plant-derived anticancer drugs such as Taxol (paclitaxel). Natural products provide diverse mechanisms of action and can be used directly or as prodrugs for further anticancer optimization. Despite the success, major bottlenecks can delay anticancer lead discovery and implementation. Recent advances in sequencing and omics-related technology have provided a mine of information for developing new therapeutics from natural products. Artificial intelligence (AI), including machine learning (ML), has offered powerful techniques for extensive data analysis and prediction-making in anticancer leads discovery. This review presents an overview of current AI-guided solutions to discover anticancer lead compounds, focusing on natural products from plants and associated microorganisms.

Molecular Modeling Applied to the Discovery of New Lead Compounds for P2 Receptors Based on Natural Sources

P2 receptors are a family of transmembrane receptors activated by nucleotides and nucleosides. Two classes have been described in mammals, P2X and P2Y, which are implicated in various diseases. Currently, only P2Y12 has medicines approved for clinical use as antiplatelet agents and natural products have emerged as a source of new drugs with action on P2 receptors due to the diversity of chemical structures. In drug discovery, in silico virtual screening (VS) techniques have become popular because they have numerous advantages, which include the evaluation of thousands of molecules against a target, usually proteins, faster and cheaper than classical high throughput screening (HTS). The number of studies using VS techniques has been growing in recent years and has led to the discovery of new molecules of natural origin with action on different P2X and P2Y receptors. Using different algorithms it is possible to obtain information on absorption, distribution, metabolism, toxicity, as well as predictions on biological activity and the lead-likeness of the selected hits. Selected biomolecules may then be tested by molecular dynamics and, if necessary, rationally designed or modified to improve their interaction for the target. The algorithms of these in silico tools are being improved to permit the precision development of new drugs and, in the future, this process will take the front of drug development against some central nervous system (CNS) disorders. Therefore, this review discusses the methodologies of in silico tools concerning P2 receptors, as well as future perspectives and discoveries, such as the employment of artificial intelligence in drug discovery.

Machine learning and AI-based approaches for bioactive ligand discovery and GPCR-ligand recognition

In the last decade, machine learning and artificial intelligence applications have received a significant boost in performance and attention in both academic research and industry. The success behind most of the recent state-of-the-art methods can be attributed to the latest developments in deep learning. When applied to various scientific domains that are concerned with the processing of non-tabular data, for example, image or text, deep learning has been shown to outperform not only conventional machine learning but also highly specialized tools developed by domain experts. This review aims to summarize AI-based research for GPCR bioactive ligand discovery with a particular focus on the most recent achievements and research trends. To make this article accessible to a broad audience of computational scientists, we provide instructive explanations of the underlying methodology, including overviews of the most commonly used deep learning architectures and feature representations of molecular data. We highlight the latest AI-based research that has led to the successful discovery of GPCR bioactive ligands. However, an equal focus of this review is on the discussion of machine learning-based technology that has been applied to ligand discovery in general and has the potential to pave the way for successful GPCR bioactive ligand discovery in the future. This review concludes with a brief outlook highlighting the recent research trends in deep learning, such as active learning and semi-supervised learning, which have great potential for advancing bioactive ligand discovery.

A Microfluidic Co-Flow Route for Human Serum Albumin-Drug-Nanoparticle Assembly

Nanoparticles are widely studied as carrier vehicles in biological systems because their size readily allows access through cellular membranes. Moreover, they have the potential to carry cargo molecules and as such, these factors make them especially attractive for intravenous drug delivery purposes. Interest in protein-based nanoparticles has recently gained attraction due to particle biocompatibility and lack of toxicity. However, the production of homogeneous protein nanoparticles with high encapsulation efficiencies, without the need for additional cross-linking or further engineering of the molecule, remains challenging. Herein, we present a microfluidic 3D co-flow device to generate human serum albumin/celastrol nanoparticles by co-flowing an aqueous protein solution with celastrol in ethanol. This microscale co-flow method resulted in the formation of nanoparticles with a homogeneous size distribution and an average size, which could be tuned from ≈100 nm to 1 μm by modulating the flow rates used. We show that the high stability of the particles stems from the covalent cross-linking of the naturally present cysteine residues within the particles formed during the assembly step. By choosing optimal flow rates during synthesis an encapsulation efficiency of 75±24 % was achieved. Finally, we show that this approach achieves significantly enhanced solubility of celastrol in the aqueous phase and, crucially, reduced cellular toxicity.

Machine learning for target discovery in drug development

The discovery of macromolecular targets for bioactive agents is currently a bottleneck for the informed design of chemical probes and drug leads. Typically, activity profiling against genetically manipulated cell lines or chemical proteomics is pursued to shed light on their biology and deconvolute drug-target networks. By taking advantage of the ever-growing wealth of publicly available bioactivity data, learning algorithms now provide an attractive means to generate statistically motivated research hypotheses and thereby prioritize biochemical screens. Here, we highlight recent successes in machine intelligence for target identification and discuss challenges and opportunities for drug discovery.