[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-126763-en":3,"doc-seo-126763-105":30,"detail-sidebar-cat-0-en-105":91},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":4,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":21,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":27,"seo_description":14,"update_tm":28,"read_time":29},126763,962084926284,"Aurora","https://ap-avatar.wpscdn.com/davatar_29158cc5080c5b710cf443261637dec0",8,"Research & Report","Deconvoluting low yield from weak potency indirect-to-biology workflows with machine learning - Research Article","High throughput biological evaluation of small molecules is a key driver in drug discovery, and direct-to-biology (D2B) speeds screening by omitting compound purification. A core limitation of D2B is that biological readouts depend on reactions yielding the intended compound; low yield and potency can create confounded false negatives. The work introduces a machine learning yield-assay deconfounder to separate low-yield effects from true biological activity. The method identifies SARS-CoV-2 main protease inhibitors with nanomolar potency comparable to standard D2B and supports broad in silico screening to recover similar potency.","RSC  \nMedicinal Chemistry  \n| RESEARCH ARTICLE  |\n| --- |\n|  |\n\nCite this: DOI: 10. 1039/d3md00719g  \nReceived 16th December 2023, Accepted 12th February 2024  \nDOI: 10.1039/d3md00719g[rsc.li/medchem](rsc.li/medchem)  \nDeconvoluting low yield from weak potency indirect-to-biology workflows with machine learning†  \nWilliam McCorkindale,a Mihajlo Filep, b Nir London, b Alpha A. Lee  a and Emma King-Smith  *c  \nHigh throughput and rapid biological evaluation of small molecules is an essential factor in drug discovery and development. Direct-to-biology (D2B), whereby compound purification is foregone, has emerged as a viable technique in time efficient screening, specifically for PROTAC design and biological evaluation. However, one notable limitation is the prerequisite of high yielding reactions to ensure the desired compound is indeed the compound responsible for biological activity. Herein, we report a machine learning based yield-assay deconfounder capable of deconvoluting low yield from low potency to identify false negatives. We validated this approach by identifying promising SARS-CoV-2 main protease inhibitors with nanomolar activity that rivaled potency observed from the standard D2B workflow. Furthermore, we show how our framework can be utilized in a broad, in silico screen to produce compounds of similar potency as a D2B assay.  \nIntroduction  \nDirect-to-biology (D2B) in combination with high throughput screening (HTS) streamlines the target discovery pipeline by obviating the need for separation and purification. Reactive chemical fragment and PROTAC design have both significantly benefitted from this style of HTS.1–4 However, D2B currently requires some careful experimental design to limit the noise of crude reactions. For example, reagent scavengers may be used to remove unwanted reaction components via filtration, but their application is limited to select reagents. For common chemistry such as peptide bond formation reactions, resins to remove residual coupling agent are commercially available, but this is not universal for all useful reactions.1 Alternatively, crude reaction mixtures can be used for biological investigation, however, the reagents utilized must not interfere with the assay to form false negatives or false positives.3 Likewise for any byproducts formed, reaction optimization should be performed to limit the number of side products.2 Often only higher yielding reactions are used in assays; lower yielding reactions are not  \na Cavendish Laboratory, University of Cambridge, UK  \nb Department of Chemical and Structural Biology, The Weizmann Institute of Science, Israel  \nc Yusuf Hamied Department of Chemistry, University of Cambridge, UK.  \nE-mail: [esk34@cam.ac.uk](esk34@cam.ac.uk)  \n† Electronic supplementary information (ESI) available. See DOI: [https://doi.org/](https://doi.org/)[ ](https://doi.org/)[10.1039/d3md00719g](10.1039/d3md00719g)  \nfurther investigated due to the confounding effects of lower yielding reactions in the biological assay. A robust method to deconvolute the inherent noise of the crude reaction from the biological activity would increase the scope of high throughput D2B to a more diverse range of chemistries.  \nMachine learning (ML) methods have been successfully applied to predicting a variety of medicinal chemistry relevant properties including QSAR, IC50 prediction of purified compounds, and improved docking studies.5–7 The principal driving force behind these computational algorithms is identifying and subsequently modelling correlations between chemical or biological features and measured outcome. Our hypothesis is that a machine learning approach can effectively act as a denoising algorithm – picking out average trends in structure–activity relationships across chemical space amid noise such as variations in yield. Prior reports utilizing ML to identify data outliers, including false negatives, have encompassed a variety of approaches in the chemical and biological area","cbCaigTXOMNCKvp0","https://ap.wps.com/l/cbCaigTXOMNCKvp0","pdf",1541086,1,7,"English","en",105,"# Introduction\n## Direct-to-biology and high throughput screening constraints\n## Confounding from low-yield crude reaction mixtures\n# Machine learning approach and hypothesis\n## Denoising via structure–activity correlations\n## Prior ML outlier and false-negative detection methods\n# Swiss cheese ML workflow (overview)","[{\"question\":\"What problem does the study address in direct-to-biology workflows?\",\"answer\":\"D2B relies on biological assays that can be confounded when crude reactions produce low-yield or low-potency mixtures, leading to false negatives. The study targets how to disentangle yield-related noise from genuine activity.\"},{\"question\":\"What is the proposed machine learning method?\",\"answer\":\"It introduces a machine learning based yield-assay deconfounder using a “Swiss cheese” approach with two distinct ML paradigms to model confounded crude reaction mixture assay data. Combined predictions help identify likely false negatives.\"},{\"question\":\"How was the approach validated and what were the results?\",\"answer\":\"The framework was validated by identifying SARS-CoV-2 main protease inhibitors exhibiting nanomolar activity that rivaled potency observed in a standard D2B workflow. It was also demonstrated as useful for broad in silico screening to yield compounds with D2B-like potency.\"}]","Deconvoluting low yield from weak potency indirect-to-biology workflows with machine learning - Research Article | PDF",1785934656,18,{"code":4,"msg":31,"data":32},"ok",{"site_id":24,"language":23,"slug":33,"title":13,"keywords":34,"description":14,"schema_data":35,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":28},"deconvoluting-low-yield-from-weak-potency-indirect-to-biology-workflows-with-machine-learning-research-article","",{"@graph":36,"@context":85},[37,54,68],{"@type":38,"itemListElement":39},"BreadcrumbList",[40,44,48,51],{"item":41,"name":42,"@type":43,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":45,"name":46,"@type":43,"position":47},"https://docshare.wps.com/document/","Document",2,{"item":49,"name":12,"@type":43,"position":50},"https://docshare.wps.com/document/research-report/",3,{"item":52,"name":13,"@type":43,"position":53},"https://docshare.wps.com/document/deconvoluting-low-yield-from-weak-potency-indirect-to-biology-workflows-with-machine-learning-research-article/126763/",4,{"url":52,"name":13,"@type":55,"author":56,"headline":13,"publisher":58,"fileFormat":61,"inLanguage":23,"description":14,"dateModified":62,"datePublished":62,"encodingFormat":61,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":57},"Person",{"url":41,"name":59,"@type":60},"DocShare","Organization","application/pdf","2026-08-05",true,{"@type":65,"interactionType":66,"userInteractionCount":4},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"What problem does the study address in direct-to-biology workflows?","Question",{"text":75,"@type":76},"D2B relies on biological assays that can be confounded when crude reactions produce low-yield or low-potency mixtures, leading to false negatives. The study targets how to disentangle yield-related noise from genuine activity.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"What is the proposed machine learning method?",{"text":80,"@type":76},"It introduces a machine learning based yield-assay deconfounder using a “Swiss cheese” approach with two distinct ML paradigms to model confounded crude reaction mixture assay data. Combined predictions help identify likely false negatives.",{"name":82,"@type":73,"acceptedAnswer":83},"How was the approach validated and what were the results?",{"text":84,"@type":76},"The framework was validated by identifying SARS-CoV-2 main protease inhibitors exhibiting nanomolar activity that rivaled potency observed in a standard D2B workflow. It was also demonstrated as useful for broad in silico screening to yield compounds with D2B-like potency.","https://schema.org",{"og:url":52,"og:type":87,"og:title":13,"og:site_name":59,"og:description":14},"article",{"robots":89,"canonical":52},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,119,122,127,130,134],{"id":20,"doc_module":4,"doc_module_name":46,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":53,"doc_module":4,"doc_module_name":46,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":46,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":46,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":21,"doc_module":4,"doc_module_name":46,"category_name":116,"show_sort_weight":117,"slug":118},"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":120,"slug":121},30,"research-report",{"id":123,"doc_module":4,"doc_module_name":46,"category_name":124,"show_sort_weight":125,"slug":126},9,"Religion & Spirituality",20,"religion-spirituality",{"id":125,"doc_module":4,"doc_module_name":46,"category_name":128,"show_sort_weight":125,"slug":129},"World Cup","world-cup",{"id":131,"doc_module":4,"doc_module_name":46,"category_name":132,"show_sort_weight":131,"slug":133},10,"Lifestyle","lifestyle",{"id":135,"doc_module":4,"doc_module_name":46,"category_name":136,"show_sort_weight":106,"slug":137},19,"General","general"]