Matches in SemOpenAlex for { <https://semopenalex.org/work/W4380885728> ?p ?o ?g. }
Showing items 1 to 71 of
71
with 100 items per page.
- W4380885728 abstract "While a multitude of deep generative models have recently emerged there exists no best practice for their practically relevant validation. On the one hand, novel de novo-generated molecules cannot be refuted by retrospective validation (so that this type of validation is biased); but on the other hand prospective validation is expensive and then often biased by the human selection process. In this case study, we frame retrospective validation as the ability to mimic human drug design, by answering the following question: Can a generative model trained on early-stage project compounds generate middle/late-stage compounds de novo? To this end, we used experimental data that contains the elapsed time of a synthetic expansion following hit identification from five public (where the time series was pre-processed to better reflect realistic synthetic expansions) and six in-house project datasets, and used REINVENT as a widely adopted RNN-based generative model. After splitting the dataset and training REINVENT on early-stage compounds, we found that rediscovery of middle/late-stage compounds was much higher in public projects (at 1.60%, 0.64%, and 0.21% of the top 100, 500, and 5,000 scored generated compounds) than in in-house projects (where the values were 0.00%, 0.03%, and 0.04%, respectively). Similarly, average single nearest neighbour similarity between early- and middle/late-stage compounds in public projects was higher between active compounds than inactive compounds; however, for in-house projects the converse was true, which makes rediscovery (if so desired) more difficult. We hence show that the generative model recovers very few middle/late-stage compounds from real-world drug discovery projects, highlighting the fundamental difference between purely algorithmic design and drug discovery as a real-world process. Evaluating de novo compound design approaches appears, based on the current study, difficult or even impossible to do retrospectively." @default.
- W4380885728 created "2023-06-17" @default.
- W4380885728 creator A5020943315 @default.
- W4380885728 creator A5026643759 @default.
- W4380885728 creator A5027609089 @default.
- W4380885728 creator A5045140841 @default.
- W4380885728 creator A5064256420 @default.
- W4380885728 date "2023-06-16" @default.
- W4380885728 modified "2023-09-25" @default.
- W4380885728 title "On The Difficulty of Validating Molecular Generative Models Realistically: A Case Study on Public and Proprietary Data" @default.
- W4380885728 doi "https://doi.org/10.26434/chemrxiv-2023-lbvgn" @default.
- W4380885728 hasPublicationYear "2023" @default.
- W4380885728 type Work @default.
- W4380885728 citedByCount "0" @default.
- W4380885728 crossrefType "posted-content" @default.
- W4380885728 hasAuthorship W4380885728A5020943315 @default.
- W4380885728 hasAuthorship W4380885728A5026643759 @default.
- W4380885728 hasAuthorship W4380885728A5027609089 @default.
- W4380885728 hasAuthorship W4380885728A5045140841 @default.
- W4380885728 hasAuthorship W4380885728A5064256420 @default.
- W4380885728 hasBestOaLocation W43808857281 @default.
- W4380885728 hasConcept C103278499 @default.
- W4380885728 hasConcept C111919701 @default.
- W4380885728 hasConcept C115961682 @default.
- W4380885728 hasConcept C119857082 @default.
- W4380885728 hasConcept C126042441 @default.
- W4380885728 hasConcept C146357865 @default.
- W4380885728 hasConcept C151730666 @default.
- W4380885728 hasConcept C154945302 @default.
- W4380885728 hasConcept C167966045 @default.
- W4380885728 hasConcept C2524010 @default.
- W4380885728 hasConcept C2776809875 @default.
- W4380885728 hasConcept C33923547 @default.
- W4380885728 hasConcept C39890363 @default.
- W4380885728 hasConcept C41008148 @default.
- W4380885728 hasConcept C76155785 @default.
- W4380885728 hasConcept C86803240 @default.
- W4380885728 hasConcept C98045186 @default.
- W4380885728 hasConceptScore W4380885728C103278499 @default.
- W4380885728 hasConceptScore W4380885728C111919701 @default.
- W4380885728 hasConceptScore W4380885728C115961682 @default.
- W4380885728 hasConceptScore W4380885728C119857082 @default.
- W4380885728 hasConceptScore W4380885728C126042441 @default.
- W4380885728 hasConceptScore W4380885728C146357865 @default.
- W4380885728 hasConceptScore W4380885728C151730666 @default.
- W4380885728 hasConceptScore W4380885728C154945302 @default.
- W4380885728 hasConceptScore W4380885728C167966045 @default.
- W4380885728 hasConceptScore W4380885728C2524010 @default.
- W4380885728 hasConceptScore W4380885728C2776809875 @default.
- W4380885728 hasConceptScore W4380885728C33923547 @default.
- W4380885728 hasConceptScore W4380885728C39890363 @default.
- W4380885728 hasConceptScore W4380885728C41008148 @default.
- W4380885728 hasConceptScore W4380885728C76155785 @default.
- W4380885728 hasConceptScore W4380885728C86803240 @default.
- W4380885728 hasConceptScore W4380885728C98045186 @default.
- W4380885728 hasLocation W43808857281 @default.
- W4380885728 hasOpenAccess W4380885728 @default.
- W4380885728 hasPrimaryLocation W43808857281 @default.
- W4380885728 hasRelatedWork W1534961803 @default.
- W4380885728 hasRelatedWork W2108501770 @default.
- W4380885728 hasRelatedWork W2164688428 @default.
- W4380885728 hasRelatedWork W2994891734 @default.
- W4380885728 hasRelatedWork W3003214776 @default.
- W4380885728 hasRelatedWork W3017062960 @default.
- W4380885728 hasRelatedWork W4241824423 @default.
- W4380885728 hasRelatedWork W4253371817 @default.
- W4380885728 hasRelatedWork W4293428270 @default.
- W4380885728 hasRelatedWork W2310403681 @default.
- W4380885728 isParatext "false" @default.
- W4380885728 isRetracted "false" @default.
- W4380885728 workType "article" @default.