Inside the mind of a medicinal chemist: the role of human bias in compound prioritization during drug discovery

DrugGym: A testbed for the economics of autonomous drug discovery

Drug discovery is stochastic. The effectiveness of candidate compounds in satisfying design objectives is unknown ahead of time, and the tools used for prioritization-predictive models and assays-are inaccurate and noisy. In a typical discovery campaign, thousands of compounds may be synthesized and tested before design objectives are achieved, with many others ideated but deprioritized. These challenges are well-documented, but assessing potential remedies has been difficult. We introduce DrugGym, a framework for modeling the stochastic process of drug discovery. Emulating biochemical assays with realistic surrogate models, we simulate the progression from weak hits to sub-micromolar leads with viable ADME. We use this testbed to examine how different ideation, scoring, and decision-making strategies impact statistical measures of utility, such as the probability of program success within predefined budgets and the expected costs to achieve target candidate profile (TCP) goals. We also assess the influence of affinity model inaccuracy, chemical creativity, batch size, and multi-step reasoning. Our findings suggest that reducing affinity model inaccuracy from 2 to 0.5 pIC50 units improves budget-constrained success rates tenfold. DrugGym represents a realistic testbed for machine learning methods applied to the hit-to-lead phase. Source code is available at www.drug-gym.org.

Molecule auto-correction to facilitate molecular design

Ensuring that computationally designed molecules are chemically reasonable is at best cumbersome. We present a molecule correction algorithm that morphs invalid molecular graphs into structurally related valid analogs. The algorithm is implemented as a tree search, guided by a set of policies to minimize its cost. We showcase how the algorithm can be applied to molecular design, either as a post-processing step or as an integral part of molecule generators.

Extracting medicinal chemistry intuition via preference machine learning

The lead optimization process in drug discovery campaigns is an arduous endeavour where the input of many medicinal chemists is weighed in order to reach a desired molecular property profile. Building the expertise to successfully drive such projects collaboratively is a very time-consuming process that typically spans many years within a chemist's career. In this work we aim to replicate this process by applying artificial intelligence learning-to-rank techniques on feedback that was obtained from 35 chemists at Novartis over the course of several months. We exemplify the usefulness of the learned proxies in routine tasks such as compound prioritization, motif rationalization, and biased de novo drug design. Annotated response data is provided, and developed models and code made available through a permissive open-source license.

A molecule perturbation software library and its application to study the effects of molecular design constraints

Computational molecular design can yield chemically unreasonable compounds when performed carelessly. A popular strategy to mitigate this risk is mimicking reference chemistry. This is commonly achieved by restricting the way in which molecules are constructed or modified. While it is well established that such an approach helps in designing chemically appealing molecules, concerns about these restrictions impacting chemical space exploration negatively linger. In this work we present a software library for constrained graph-based molecule manipulation and showcase its functionality by developing a molecule generator. Said generator designs molecules mimicking reference chemical features of differing granularity. We find that restricting molecular construction lightly, beyond the usual positive effects on drug-likeness and synthesizability of designed molecules, provides guidance to optimization algorithms navigating chemical space. Nonetheless, restricting molecular construction excessively can indeed hinder effective chemical space exploration.

Impact of Applicability Domains to Generative Artificial Intelligence

Molecular generative artificial intelligence is drawing significant attention in the drug design community, with several experimentally validated proof of concepts already published. Nevertheless, generative models are known for sometimes generating unrealistic, unstable, unsynthesizable, or uninteresting structures. This calls for methods to constrain those algorithms to generate structures in drug-like portions of the chemical space. While the concept of applicability domains for predictive models is well studied, its counterpart for generative models is not yet well-defined. In this work, we empirically examine various possibilities and propose applicability domains suited for generative models. Using both public and internal data sets, we use generative methods to generate novel structures that are predicted to be actives by a corresponding quantitative structure-activity relationships model while constraining the generative model to stay within a given applicability domain. Our work looks at several applicability domain definitions, combining various criteria, such as structural similarity to the training set, similarity of physicochemical properties, unwanted substructures, and quantitative estimate of drug-likeness. We assess the structures generated from both qualitative and quantitative points of view and find that the applicability domain definitions have a strong influence on the drug-likeness of generated molecules. An extensive analysis of our results allows us to identify applicability domain definitions that are best suited for generating drug-like molecules with generative models. We anticipate that this work will help foster the adoption of generative models in an industrial context.

G2Retro as a two-step graph generative models for retrosynthesis prediction

Retrosynthesis is a procedure where a target molecule is transformed into potential reactants and thus the synthesis routes can be identified. Recently, computational approaches have been developed to accelerate the design of synthesis routes. In this paper,we develop a generative framework G²Retro for one-step retrosynthesis prediction. G²Retro imitates the reversed logic of synthetic reactions. It first predicts the reaction centers in the target molecules (products), identifies the synthons needed to assemble the products, and transforms these synthons into reactants. G²Retro defines a comprehensive set of reaction center types, and learns from the molecular graphs of the products to predict potential reaction centers. To complete synthons into reactants, G²Retro considers all the involved synthon structures and the product structures to identify the optimal completion paths, and accordingly attaches small substructures sequentially to the synthons. Here we show that G²Retro is able to better predict the reactants for given products in the benchmark dataset than the state-of-the-art methods.

Quantitative evaluation of explainable graph neural networks for molecular property prediction

Graph neural networks (GNNs) have received increasing attention because of their expressive power on topological data, but they are still criticized for their lack of interpretability. To interpret GNN models, explainable artificial intelligence (XAI) methods have been developed. However, these methods are limited to qualitative analyses without quantitative assessments from the real-world datasets due to a lack of ground truths. In this study, we have established five XAI-specific molecular property benchmarks, including two synthetic and three experimental datasets. Through the datasets, we quantitatively assessed six XAI methods on four GNN models and made comparisons with seven medicinal chemists of different experience levels. The results demonstrated that XAI methods could deliver reliable and informative answers for medicinal chemists in identifying the key substructures. Moreover, the identified substructures were shown to complement existing classical fingerprints to improve molecular property predictions, and the improvements increased with the growth of training data.

Molecular Similarity Perception Based on Machine-Learning Models

Molecular similarity is an impressively broad topic with many implications in several areas of chemistry. Its roots lie in the paradigm that ‘similar molecules have similar properties’. For this reason, methods for determining molecular similarity find wide application in pharmaceutical companies, e.g., in the context of structure-activity relationships. The similarity evaluation is also used in the field of chemical legislation, specifically in the procedure to judge if a new molecule can obtain the status of orphan drug with the consequent financial benefits. For this procedure, the European Medicines Agency uses experts’ judgments. It is clear that the perception of the similarity depends on the observer, so the development of models to reproduce the human perception is useful. In this paper, we built models using both 2D fingerprints and 3D descriptors, i.e., molecular shape and pharmacophore descriptors. The proposed models were also evaluated by constructing a dataset of pairs of molecules which was submitted to a group of experts for the similarity judgment. The proposed machine-learning models can be useful to reduce or assist human efforts in future evaluations. For this reason, the new molecules dataset and an online tool for molecular similarity estimation have been made freely available.

Automation and data-driven design of polymer therapeutics

Polymers are uniquely suited for drug delivery and biomaterial applications due to tunable structural parameters such as length, composition, architecture, and valency. To facilitate designs, researchers may explore combinatorial libraries in a high throughput fashion to correlate structure to function. However, traditional polymerization reactions including controlled living radical polymerization (CLRP) and ring-opening polymerization (ROP) require inert reaction conditions and extensive expertise to implement. With the advent of air-tolerance and automation, several polymerization techniques are now compatible with well plates and can be carried out at the benchtop, making high throughput synthesis and high throughput screening (HTS) possible. To avoid HTS pitfalls often described as "fishing expeditions," it is crucial to employ intelligent and big data approaches to maximize experimental efficiency. This is where the disruptive technologies of machine learning (ML) and artificial intelligence (AI) will likely play a role. In fact, ML and AI are already impacting small molecule drug discovery and showing signs of emerging in drug delivery. In this review, we present state-of-the-art research in drug delivery, gene delivery, antimicrobial polymers, and bioactive polymers alongside data-driven developments in drug design and organic synthesis. From this insight, important lessons are revealed for the polymer therapeutics community including the value of a closed loop design-build-test-learn workflow. This is an exciting time as researchers will gain the ability to fully explore the polymer structural landscape and establish quantitative structure-property relationships (QSPRs) with biological significance.

Evolution of Novartis' Small Molecule Screening Deck Design

This article summarizes the evolution of the screening deck at the Novartis Institutes for BioMedical Research (NIBR). Historically, the screening deck was an assembly of all available compounds. In 2015, we designed a first deck to facilitate access to diverse subsets with optimized properties. We allocated the compounds as plated subsets on a 2D grid with property based ranking in one dimension and increasing structural redundancy in the other. The learnings from the 2015 screening deck were applied to the design of a next generation in 2019. We found that using traditional leadlikeness criteria (mainly MW, clogP) reduces the hit rates of attractive chemical starting points in subset screening. Consequently, the 2019 deck relies on solubility and permeability to select preferred compounds. The 2019 design also uses NIBR's experimental assay data and inferred biological activity profiles in addition to structural diversity to define redundancy across the compound sets.

Learning Molecular Representations for Medicinal Chemistry

The accurate modeling and prediction of small molecule properties and bioactivities depend on the critical choice of molecular representation. Decades of informatics-driven research have relied on expert-designed molecular descriptors to establish quantitative structure-activity and structure-property relationships for drug discovery. Now, advances in deep learning make it possible to efficiently and compactly learn molecular representations directly from data. In this review, we discuss how active research in molecular deep learning can address limitations of current descriptors and fingerprints while creating new opportunities in cheminformatics and virtual screening. We provide a concise overview of the role of representations in cheminformatics, key concepts in deep learning, and argue that learning representations provides a way forward to improve the predictive modeling of small molecule bioactivities and properties.

Matched Molecular Series Analysis for ADME Property Prediction

Generation and prioritization of new molecules are the most central part of the drug design process. Matched molecular series analysis (MMSA) has recently been proposed as a formal approach that captures both of these key elements of design. In order to better understand the power of MMSA and its specific limitations, we here evaluate its performance as an ADME property prediction tool. We use four large and diverse inhouse data sets, logD, microsomal clearance, CYP2C9, and CYP3A4 inhibition. MMSA follows the concept of parallel structure-activity relationship (SAR), where if two identical substituent series on different scaffolds show similarity in their property profiles, SAR from one series can be transferred to the other series. We test four different similarity metrics to identify pairs of molecular series where information can be transferred. We find that the best prediction performance is achieved by a combination of centered root-mean-square deviation (cRMSD) and a network score approach previously published by Keefer et al. However, cRMSD alone strikes the best balance between accuracy and the number of predictions that can be made. We identify statistical metrics that allow estimating when MMSA predictions will work, similar to the well-known applicability domain concept in machine learning. MMSA achieves a prediction accuracy that is comparable to a standard machine-learning model and matched molecular pair analysis. In contrast to machine learning, however, it is very easy to understand where MMSA predictions are coming from. Finally, to prospectively test the power of MMSA, we retested compounds that were strong outliers in the initial predictions and show how the MMSA model can help to identify erroneous data points.

Diversity & tractability revisited in collaborative small molecule phenotypic screening library design

Identification of purposeful chemical matter on a broad range of drug targets is of high importance to the pharmaceutical industry. However, disease-relevant but more complex hit-finding plans require flexibility regarding the subset of the compounds that we screen. Herein we describe a strategy to design high-quality small molecule screening subsets of two different sizes to cope with a rapidly changing early discovery portfolio. The approach taken balances chemical tractability, chemical diversity and biological target coverage. Furthermore, using surveys, we actively involved chemists within our company in the selection process of the diversity decks to ensure current medicinal chemistry principles were incorporated. The chemist surveys revealed that not all published PAINS substructure alerts are considered productive by the medicinal chemistry community and in agreement with previously published results from other institutions, QED scores tracked quite well with chemists' notions of chemical attractiveness.

Targeting RNA with Small Molecules: Identification of Selective, RNA-Binding Small Molecules Occupying Drug-Like Chemical Space

Although the potential value of RNA as a target for new small molecule therapeutics is becoming increasingly credible, the physicochemical properties required for small molecules to selectively bind to RNA remain relatively unexplored. To investigate the druggability of RNAs with small molecules, we have employed affinity mass spectrometry, using the Automated Ligand Identification System (ALIS), to screen 42 RNAs from a variety of RNA classes, each against an array of chemically diverse drug-like small molecules (~50,000 compounds) and functionally annotated tool compounds (~5100 compounds). The set of RNA–small molecule interactions that was generated was compared with that for protein–small molecule interactions, and naïve Bayesian models were constructed to determine the types of specific chemical properties that bias small molecules toward binding to RNA. This set of RNA-selective chemical features was then used to build an RNA-focused set of ~3800 small molecules that demonstrated increased propensity toward binding the RNA target set. In addition, the data provide an overview of the specific physicochemical properties that help to enable binding to potential RNA targets. This work has increased the understanding of the chemical properties that are involved in small molecule binding to RNA, and the methodology used here is generally applicable to RNA-focused drug discovery efforts.

Decision Making in Medicinal Chemistry: The Power of Our Intuition

Medicinal chemists rely on their intuition to make decisions regarding the course of a medicinal chemistry program. Our ability to accurately and efficiently process large data sets routinely requires that we reduce the volume of information to manageable proportions. This prioritization process, however, can be affected by intuitive biases. One such situation is structure-activity relationship (SAR) analysis in nonadditive data sets in which attempts to intuitively predict the activity of compounds based on preliminary data can lead to erroneous conclusions. Matrix analysis can be a useful tool to accurately determine the nature of the SAR and to improve our decision-making process during an analoging campaign.

Wisdom of crowds for synthetic accessibility evaluation

Synthetic accessibility evaluation is a process to assess the ease of synthesis of compounds. A rapid method for the assessment of synthetic accessibility for a vast number of chemical compounds is expected to bring about a breakthrough in the drug discovery. Although several computational methods have been proposed, the compound evaluation has still been processed by medicinal chemists; however, the low throughput of the human evaluation due to the lack of chemists is a critical issue for handling a large number of compounds. We propose the use of crowdsourcing for addressing this problem, and we conducted experiments to investigate the feasibility of incorporating semi-experts and a statistical aggregation method into the synthetic accessibility evaluation. Our experimental results show that we can obtain accurate synthetic accessibility scores through the statistical aggregation of judgments from semi-experts.

Nonpher: computational method for design of hard-to-synthesize structures

In cheminformatics, machine learning methods are typically used to classify chemical compounds into distinctive classes such as active/nonactive or toxic/nontoxic. To train a classifier, a training data set must consist of examples from both positive and negative classes. While a biological activity or toxicity can be experimentally measured, another important molecular property, a synthetic feasibility, is a more abstract feature that can’t be easily assessed. In the present paper, we introduce Nonpher, a computational method for the construction of a hard-to-synthesize virtual library. Nonpher is based on a molecular morphing algorithm in which new structures are iteratively generated by simple structural changes, such as the addition or removal of an atom or a bond. In Nonpher, molecular morphing was optimized so that it yields structures not overly complex, but just right hard-to-synthesize. Nonpher results were compared with SAscore and dense region (DR), other two methods for the generation of hard-to-synthesize compounds. Random forest classifier trained on Nonpher data achieves better results than models obtained using SAscore and DR data.

mQC: A Heuristic Quality-Control Metric for High-Throughput Drug Combination Screening

Quality control (QC) metrics are critical in high throughput screening (HTS) platforms to ensure reliability and confidence in assay data and downstream analyses. Most reported HTS QC metrics are designed for plate level or single well level analysis. With the advent of high throughput combination screening there is a need for QC metrics that quantify the quality of combination response matrices. We introduce a predictive, interpretable, matrix-level QC metric, mQC, based on a mix of data-derived and heuristic features. mQC accurately reproduces the expert assessment of combination response quality and correctly identifies unreliable response matrices that can lead to erroneous or misleading characterization of synergy. When combined with the plate-level QC metric, Z', mQC provides a more appropriate determination of the quality of a drug combination screen. Retrospective analysis on a number of completed combination screens further shows that mQC is able to identify problematic screens whereas plate-level QC was not able to. In conclusion, our data indicates that mQC is a reliable QC filter that can be used to identify problematic drug combinations matrices and prevent further analysis on erroneously active combinations as well as for troubleshooting failed screens. The R source code of mQC is available at http://matrix.ncats.nih.gov/mQC.

Fragment library design: using cheminformatics and expert chemists to fill gaps in existing fragment libraries

Fragment based screening (FBS) has emerged as a mainstream lead discovery strategy in academia, biotechnology start-ups, and large pharma. As a prerequisite of FBS, a structurally diverse library of fragments is desirable in order to identify chemical matter that will interact with the range of diverse target classes that are prosecuted in contemporary screening campaigns. In addition, it is also desirable to offer synthetically amenable starting points to increase the probability of a successful fragment evolution through medicinal chemistry. Herein we describe a method to identify biologically relevant chemical substructures that are missing from an existing fragment library (chemical gaps), and organize these chemical gaps hierarchically so that medicinal chemists can efficiently navigate the prioritized chemical space and subsequently select purchasable fragments for inclusion in an enhanced fragment library.

VSViewer3D: a tool for interactive data mining of three-dimensional virtual screening data

The VSviewer3D is a simple Java tool for visual exploration of three-dimensional (3D) virtual screening data. The VSviewer3D brings together the ability to explore numerical data, such as calculated properties and virtual screening scores, structure depiction, interactive topological and 3D similarity searching, and 3D visualization. By doing so the user is better able to quickly identify outliers, assess tractability of large numbers of compounds, visualize hits of interest, annotate hits, and mix and match interesting scaffolds. We demonstrate the utility of the VSviewer3D by describing a use case in a docking based virtual screen.

Adamantyl analogues of paracetamol as potent analgesic drugs via inhibition of TRPA1

Paracetamol also known as acetaminophen, is a widely used analgesic and antipyretic agent. We report the synthesis and biological evaluation of adamantyl analogues of paracetamol with important analgesic properties. The mechanism of nociception of compound 6a/b, an analog of paracetamol, is not exerted through direct interaction with cannabinoid receptors, nor by inhibiting COX. It behaves as an interesting selective TRPA1 channel antagonist, which may be responsible for its analgesic properties, whereas it has no effect on the TRPM8 nor TRPV1 channels. The possibility of replacing a phenyl ring by an adamantyl ring opens new avenues in other fields of medicinal chemistry.

Computational prediction and validation of an expert's evaluation of chemical probes

In a decade with over half a billion dollars of investment, more than 300 chemical probes have been identified to have biological activity through NIH funded screening efforts. We have collected the evaluations of an experienced medicinal chemist on the likely chemistry quality of these probes based on a number of criteria including literature related to the probe and potential chemical reactivity. Over 20% of these probes were found to be undesirable. Analysis of the molecular properties of these compounds scored as desirable suggested higher pKa, molecular weight, heavy atom count, and rotatable bond number. We were particularly interested whether the human evaluation aspect of medicinal chemistry due diligence could be computationally predicted. We used a process of sequential Bayesian model building and iterative testing as we included additional probes. Following external validation of these methods and comparing different machine learning methods, we identified Bayesian models with accuracy comparable to other measures of drug-likeness and filtering rules created to date.

The essential roles of chemistry in high-throughput screening triage

It is increasingly clear that academic high-throughput screening (HTS) and virtual HTS triage suffers from a lack of scientists trained in the art and science of early drug discovery chemistry. Many recent publications report the discovery of compounds by screening that are most likely artifacts or promiscuous bioactive compounds, and these results are not placed into the context of previous studies. For HTS to be most successful, it is our contention that there must exist an early partnership between biologists and medicinal chemists. Their combined skill sets are necessary to design robust assays and efficient workflows that will weed out assay artifacts, false positives, promiscuous bioactive compounds and intractable screening hits, efforts that ultimately give projects a better chance at identifying truly useful chemical matter. Expertise in medicinal chemistry, cheminformatics and purification sciences (analytical chemistry) can enhance the post-HTS triage process by quickly removing these problematic chemotypes from consideration, while simultaneously prioritizing the more promising chemical matter for follow-up testing. It is only when biologists and chemists collaborate effectively that HTS can manifest its full promise.

The use of 2D fingerprint methods to support the assessment of structural similarity in orphan drug legislation

BACKGROUND

In the European Union, medicines are authorised for some rare disease only if they are judged to be dissimilar to authorised orphan drugs for that disease. This paper describes the use of 2D fingerprints to show the extent of the relationship between computed levels of structural similarity for pairs of molecules and expert judgments of the similarities of those pairs. The resulting relationship can be used to provide input to the assessment of new active compounds for which orphan drug authorisation is being sought.

RESULTS

143 experts provided judgments of the similarity or dissimilarity of 100 pairs of drug-like molecules from the DrugBank 3.0 database. The similarities of these pairs were also computed using BCI, Daylight, ECFC4, ECFP4, MDL and Unity 2D fingerprints. Logistic regression analyses demonstrated a strong relationship between the human and computed similarity assessments, with the resulting regression models having significant predictive power in experiments using data from submissions of orphan drug medicines to the European Medicines Agency. The BCI fingerprints performed best overall on the DrugBank dataset while the BCI, Daylight, ECFP4 and Unity fingerprints performed comparably on the European Medicines Agency dataset.

CONCLUSIONS

Measures of structural similarity based on 2D fingerprints can provide a useful source of information for the assessment of orphan drug status by regulatory authorities.

Chemical predictive modelling to improve compound quality

The 'quality' of small-molecule drug candidates, encompassing aspects including their potency, selectivity and ADMET (absorption, distribution, metabolism, excretion and toxicity) characteristics, is a key factor influencing the chances of success in clinical trials. Importantly, such characteristics are under the control of chemists during the identification and optimization of lead compounds. Here, we discuss the application of computational methods, particularly quantitative structure-activity relationships (QSARs), in guiding the selection of higher-quality drug candidates, as well as cultural factors that may have affected their use and impact.

Comparative mode-of-action analysis following manual and automated phenotype detection in Xenopus laevis

Given the increasing utilization of phenotypic screens in drug discovery also the subsequent mechanism-of-action analysis gains increased attention.

Anthropogenic reaction parameters – the missing link between chemical intuition and the available chemical space

Anthropogenic factors limit reaction parameters and thus the scope of synthetic chemistry, nevertheless, their role is both advantageous and critical.

Data-driven medicinal chemistry in the era of big data

Science, and the way we undertake research, is changing. The increasing rate of data generation across all scientific disciplines is providing incredible opportunities for data-driven research, with the potential to transform our current practices. The exploitation of so-called 'big data' will enable us to undertake research projects never previously possible but should also stimulate a re-evaluation of all our data practices. Data-driven medicinal chemistry approaches have the potential to improve decision making in drug discovery projects, providing that all researchers embrace the role of 'data scientist' and uncover the meaningful relationships and patterns in available data.

How diverse are diversity assessment methods? A comparative analysis and benchmarking of molecular descriptor space

Chemical diversity is a widely applied approach to select structurally diverse subsets of molecules, often with the objective of maximizing the number of hits in biological screening. While many methods exist in the area, few systematic comparisons using current descriptors in particular with the objective of assessing diversity in bioactivity space have been published, and this shortage is what the current study is aiming to address. In this work, 13 widely used molecular descriptors were compared, including fingerprint-based descriptors (ECFP4, FCFP4, MACCS keys), pharmacophore-based descriptors (TAT, TAD, TGT, TGD, GpiDAPH3), shape-based descriptors (rapid overlay of chemical structures (ROCS) and principal moments of inertia (PMI)), a connectivity-matrix-based descriptor (BCUT), physicochemical-property-based descriptors (prop2D), and a more recently introduced molecular descriptor type (namely, "Bayes Affinity Fingerprints"). We assessed both the similar behavior of the descriptors in assessing the diversity of chemical libraries, and their ability to select compounds from libraries that are diverse in bioactivity space, which is a property of much practical relevance in screening library design. This is particularly evident, given that many future targets to be screened are not known in advance, but that the library should still maximize the likelihood of containing bioactive matter also for future screening campaigns. Overall, our results showed that descriptors based on atom topology (i.e., fingerprint-based descriptors and pharmacophore-based descriptors) correlate well in rank-ordering compounds, both within and between descriptor types. On the other hand, shape-based descriptors such as ROCS and PMI showed weak correlation with the other descriptors utilized in this study, demonstrating significantly different behavior. We then applied eight of the molecular descriptors compared in this study to sample a diverse subset of sample compounds (4%) from an initial population of 2587 compounds, covering the 25 largest human activity classes from ChEMBL and measured the coverage of activity classes by the subsets. Here, it was found that "Bayes Affinity Fingerprints" achieved an average coverage of 92% of activity classes. Using the descriptors ECFP4, GpiDAPH3, TGT, and random sampling, 91%, 84%, 84%, and 84% of the activity classes were represented in the selected compounds respectively, followed by BCUT, prop2D, MACCS, and PMI (in order of decreasing performance). In addition, we were able to show that there is no visible correlation between compound diversity in PMI space and in bioactivity space, despite frequent utilization of PMI plots to this end. To summarize, in this work, we assessed which descriptors select compounds with high coverage of bioactivity space, and can hence be used for diverse compound selection for biological screening. In cases where multiple descriptors are to be used for diversity selection, this work describes which descriptors behave complementarily, and can hence be used jointly to focus on different aspects of diversity in chemical space.

Molecular similarity in medicinal chemistry

Similarity is a subjective and multifaceted concept, regardless of whether compounds or any other objects are considered. Despite its intrinsically subjective nature, attempts to quantify the similarity of compounds have a long history in chemical informatics and drug discovery. Many computational methods employ similarity measures to identify new compounds for pharmaceutical research. However, chemoinformaticians and medicinal chemists typically perceive similarity in different ways. Similarity methods and numerical readouts of similarity calculations are probably among the most misunderstood computational approaches in medicinal chemistry. Herein, we evaluate different similarity concepts, highlight key aspects of molecular similarity analysis, and address some potential misunderstandings. In addition, a number of practical aspects concerning similarity calculations are discussed.