Evaluating the performance of drug-repurposing technologies

Schuler, James; Falls, Zackary; Mangione, William; Hudson, Matthew L; Bruggemann, Liana; Samudrala, Ram

doi:10.1016/j.drudis.2021.08.002

Cited by 26 publications

(52 citation statements)

References 131 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We also compared our designs to benchmarks from GuacaMol [60]. These results corroborated the outputs of our benchmarking and/or SCScore; future work will include a rigorous evaluation of drug design technologies, much as we have done for repurposing [29].…”

Section: Synthetic Feasibility Of Designed Compoundssupporting

confidence: 66%

“…A critical aspect of verifying the utility of the designs generated was to compare their predicted behavior (i.e., proteomic interaction signatures) to their intended behavior, which was input to conditional generation. If the predicted behaviors of designed compounds were highly similar to the conditional objective across objectives relative to the corresponding controls, we concluded that the RCVAE design pipeline may be used to accurately design compounds that possess any desirable bioactivity and subsequent function, given the extensive benchmarking and validation the CANDO paradigm has undergone [18,21,26,29,[47][48][49][50]. This is the primary motivation and goal for using the CVAE architecture in terms of accelerating drug discovery: design with respect to arbitrary numbers of on-, off-, and anti-targets (Figure 1).…”

Section: Behavioral Similarity Of Designed Compounds To Their Objectivesmentioning

confidence: 99%

“…Computational methods are efficient, accurate, holistic (i.e., take into account the entire interaction space of chemical entities), and have breadth in terms of chemical space exploration necessary to overcome the limitations of traditional approaches [2,6,12,13,[17][18][19][20][21][22][23][24][25][26][27][28][29][30][31][32][33][34]. To expand compound libraries utilized in screening, combinatorial chemistry and machine-learning design pipelines have been developed to generate libraries of compounds likely to bind to a given target [35][36][37].…”

Section: Introductionmentioning

confidence: 99%

“…We developed the Computational Analysis of Novel Drug Repurposing Opportunities (CANDO) platform for shotgun multitarget drug discovery, repurposing, and design to overcome the aforementioned limitations of traditional single-target approaches [18][19][20][21][22][23][24][25][26][27][28][29]. The platform screens and ranks drugs/compounds for every disease/indication (and adverse event) through the large-scale modeling and analytics of the interactions between comprehensive libraries of drugs/compounds and protein structures.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

A Deep-Learning Proteomic-Scale Approach for Drug Design

et al. 2021

Self Cite

View full text Add to dashboard Cite

Computational approaches have accelerated novel therapeutic discovery in recent decades. The Computational Analysis of Novel Drug Opportunities (CANDO) platform for shotgun multitarget therapeutic discovery, repurposing, and design aims to improve their efficacy and safety by employing a holistic approach that computes interaction signatures between every drug/compound and a large library of non-redundant protein structures corresponding to the human proteome fold space. These signatures are compared and analyzed to determine if a given drug/compound is efficacious and safe for a given indication/disease. In this study, we used a deep learning-based autoencoder to first reduce the dimensionality of CANDO-computed drug–proteome interaction signatures. We then employed a reduced conditional variational autoencoder to generate novel drug-like compounds when given a target encoded “objective” signature. Using this approach, we designed compounds to recreate the interaction signatures for twenty approved and experimental drugs and showed that 16/20 designed compounds were predicted to be significantly (p-value ≤ 0.05) more behaviorally similar relative to all corresponding controls, and 20/20 were predicted to be more behaviorally similar relative to a random control. We further observed that redesigns of objectives developed via rational drug design performed significantly better than those derived from natural sources (p-value ≤ 0.05), suggesting that the model learned an abstraction of rational drug design. We also show that the designed compounds are structurally diverse and synthetically feasible when compared to their respective objective drugs despite consistently high predicted behavioral similarity. Finally, we generated new designs that enhanced thirteen drugs/compounds associated with non-small cell lung cancer and anti-aging properties using their predicted proteomic interaction signatures. his study represents a significant step forward in automating holistic therapeutic design with machine learning, enabling the rapid generation of novel, effective, and safe drug leads for any indication.

show abstract

Section: Synthetic Feasibility Of Designed Compoundssupporting

confidence: 66%

Section: Behavioral Similarity Of Designed Compounds To Their Objectivesmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

A Deep-Learning Proteomic-Scale Approach for Drug Design

et al. 2021

Self Cite

View full text Add to dashboard Cite

show abstract

“…For trials which are deemed to be in-scope, the following information is recorded: In addition to the online version, a tab-delimited values text version of the database is also available for download so that the data can be analysed off-line or included in data pipelines such as BioDWH2 (Friedrichs, 2021), CANDO (Schuler et al, 2021) and other bioinformatics platforms.…”

Section: Methodsmentioning

confidence: 99%