The opportunities of mining historical and collective data in drug discovery

Wassermann, Anne Mai; Lounkine, Eugen; Davies, John W.; Glick, Meir; Camargo, Luiz Miguel

doi:10.1016/j.drudis.2014.11.004

Cited by 32 publications

(21 citation statements)

References 95 publications

(112 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The current version of the CC is organized in 5 levels of complexity (A: Chemistry, B: Targets, C: Networks, D: Cells and E: Clinics), each of which is divided into 5 sublevels (1)(2)(3)(4)(5). In total, the CC is composed of 25 spaces capturing the 2D/3D structures of the molecules, targets and metabolic genes, network properties of the targets, cell response profiles, drug indications and side effects, among others ( Figure 1a).…”

Section: Resultsmentioning

confidence: 99%

“…The corpus of bioactivity records available suggests that other numerical representations of molecules are possible, reaching beyond chemical structures and capturing their known biological properties. Indeed, it has been shown that an enriched representation of molecules can be achieved through the use of 'bioactivity signatures' 3 . Bioactivity signatures are multi-dimensional vectors that capture the biological traits of the molecule (for example, its target profile) in a format that is akin to the structural descriptors or fingerprints used in the field of chemoinformatics.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Bioactivity descriptors for uncharacterized compounds

Bertoni

Duran‐Frigola

Badia-i-Mompel

et al. 2020

Preprint

View full text Add to dashboard Cite

Chemical descriptors encode the physicochemical and structural properties of small molecules, and they are at the core of chemoinformatics. The broad release of bioactivity data has prompted enriched representations of compounds, reaching beyond chemical structures and capturing their known biological properties. Unfortunately, ‘bioactivity descriptors’ are not available for most small molecules, which limits their applicability to a few thousand well characterized compounds. Here we present a collection of deep neural networks able to infer bioactivity signatures for any compound of interest, even when little or no experimental information is available for them. Our ‘signaturizers’ relate to bioactivities of 25 different types (including target profiles, cellular response and clinical outcomes) and can be used as drop-in replacements for chemical descriptors in day-to-day chemoinformatics tasks. Indeed, we illustrate how inferred bioactivity signatures are useful to navigate the chemical space in a biologically relevant manner, and unveil higher-order organization in drugs and natural product collections. Moreover, we implement a battery of signature-activity relationship (SigAR) models and show a substantial improvement in performance, with respect to chemistry-based classifiers, across a series of biophysics and physiology activity prediction benchmarks.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Bioactivity descriptors for uncharacterized compounds

Bertoni

Duran‐Frigola

Badia-i-Mompel

et al. 2020

Preprint

View full text Add to dashboard Cite

show abstract

“…[81] Recently, these approaches have been merged and further improved by simultaneously analyzing the behavior of compounds across multiple assays and clustering assays based on the compounds that they pick up. [82] Comparing the two axes of this compound-assay matrix seems a promising way to distinguish between common assay artifacts or general non-stoichiometric inhibitors affecting multiple related assays on the one hand and genuine polypharmacology based on stoichiometric inhibition on the other hand.…”

Section: In Silico Analysismentioning

confidence: 99%

Non-stoichiometric inhibition in integrated lead finding – a literature review

Klumpp

2015

Expert Opinion on Drug Discovery

View full text Add to dashboard Cite

Over the last few years, awareness of non-stoichiometric inhibitors within screening libraries and HTS hit lists has considerably increased, not only in the pharmaceutical industry but also in the academic drug discovery community. This has resulted in a variety of methods to detect and handle such compounds. These range from in silico approaches to flag suspicious compounds, and counterassays to measure non-stoichiometric inhibition, to biophysical methods that positively demonstrate stoichiometric binding. In addition, novel technologies to verify target engagement within cells are becoming available. While still a time- and resource-consuming nuisance, non-stoichiometric inhibitors therefore do not fundamentally jeopardize the discovery of low molecular weight lead and drug candidates. Rather, they should be viewed as a manageable issue that with appropriate expertise can be overcome through integration of the above-mentioned approaches.

show abstract

“…Beyond the in-house bioactivity data collections generated and stored by pharmaceutical companies, several initiatives such as ChEMBL have made bioactivity data for millions of compounds publicly available (Gaulton et al, 2012). These range from interpretable classifiers such as inductive rules and decision trees (Drakakis, Moledina, Chomenidis, & Doganis, 2016) all the way to neural networks (Koutsoukas, Monaghan, Li, & Huan, 2017;Lenselink et al, 2017;Ma, Sheridan, Liaw, Dahl, & Svetnik, 2015;Ramsundar et al, 2015), random forests (Mervin et al, 2015), and support-vector machines (Wassermann, Lounkine, Davies, Glick, & Camargo, 2015). Target prediction refers to the algorithms used in cheminformatics to infer the mechanisms of action of small molecules by predicting their most likely protein targets from, e.g., their chemical structure (Koutsoukas et al, 2011) or by comparing the bioactivity profile of a test compound against those with known MoA across, e.g., a cancer cell line panel (Shoemaker, 2006).…”

Section: Target Predictionmentioning

confidence: 99%

Elucidating Compound Mechanism of Action and Predicting Cytotoxicity Using Machine Learning Approaches, Taking Prediction Confidence into Account

Drakakis

Cortés-Ciriano

Alexander-Dann

et al. 2019

CP Chemical Biology

View full text Add to dashboard Cite

The modes of action (MoAs) of drugs frequently are unknown, because many are small molecules initially identified from phenotypic screens, giving rise to the need to elucidate their MoAs. In addition, the high attrition rate for candidate drugs in preclinical studies due to intolerable toxicity has motivated the development of computational approaches to predict drug candidate (cyto)toxicity as early as possible in the drug-discovery process. Here, we provide detailed instructions for capitalizing on bioactivity predictions to elucidate the MoAs of small molecules and infer their underlying phenotypic effects. We illustrate how these predictions can be used to infer the underlying antidepressive effects of marketed drugs. We also provide the necessary functionalities to model cytotoxicity data using single and ensemble machine-learning algorithms. Finally, we give detailed instructions on how to calculate confidence intervals for individual predictions using the conformal prediction framework. C 2019 by John Wiley & Sons, Inc.Keywords: ChEMBL r cytotoxicity r in silico bioactivity prediction r mechanism of action r polypharmacology r toxicology modeling

show abstract

The opportunities of mining historical and collective data in drug discovery

Cited by 32 publications

References 95 publications

Bioactivity descriptors for uncharacterized compounds

Bioactivity descriptors for uncharacterized compounds

Non-stoichiometric inhibition in integrated lead finding – a literature review

Elucidating Compound Mechanism of Action and Predicting Cytotoxicity Using Machine Learning Approaches, Taking Prediction Confidence into Account

Contact Info

Product

Resources

About