Publications
16021941
4WTWCSN6
1
apa
50
date
desc
year
1
249
https://airfoundry.upenn.edu/wp-content/plugins/zotpress/
%7B%22status%22%3A%22success%22%2C%22updateneeded%22%3Afalse%2C%22instance%22%3Afalse%2C%22meta%22%3A%7B%22request_last%22%3A0%2C%22request_next%22%3A0%2C%22used_cache%22%3Atrue%7D%2C%22data%22%3A%5B%7B%22key%22%3A%22CP2ZA38S%22%2C%22library%22%3A%7B%22id%22%3A16021941%7D%2C%22meta%22%3A%7B%22creatorSummary%22%3A%22Torres%20et%20al.%22%2C%22parsedDate%22%3A%222024-11-27%22%2C%22numChildren%22%3A1%7D%2C%22bib%22%3A%22%3Cdiv%20class%3D%5C%22csl-bib-body%5C%22%20style%3D%5C%22line-height%3A%202%3B%20padding-left%3A%201em%3B%20text-indent%3A-1em%3B%5C%22%3E%5Cn%20%20%3Cdiv%20class%3D%5C%22csl-entry%5C%22%3ETorres%2C%20M.%20D.%20T.%2C%20Zeng%2C%20Y.%2C%20Wan%2C%20F.%2C%20Maus%2C%20N.%2C%20Gardner%2C%20J.%2C%20%26amp%3B%20Fuente-Nunez%2C%20C.%20de%20la.%20%282024%29.%20%3Ci%3EA%20generative%20artificial%20intelligence%20approach%20for%20antibiotic%20optimization%3C%5C%2Fi%3E.%20bioRxiv.%20%3Ca%20class%3D%27zp-DOIURL%27%20href%3D%27https%3A%5C%2F%5C%2Fdoi.org%5C%2F10.1101%5C%2F2024.11.27.625757%27%3Ehttps%3A%5C%2F%5C%2Fdoi.org%5C%2F10.1101%5C%2F2024.11.27.625757%3C%5C%2Fa%3E%3C%5C%2Fdiv%3E%5Cn%3C%5C%2Fdiv%3E%22%2C%22data%22%3A%7B%22itemType%22%3A%22preprint%22%2C%22title%22%3A%22A%20generative%20artificial%20intelligence%20approach%20for%20antibiotic%20optimization%22%2C%22creators%22%3A%5B%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Marcelo%20D.%20T.%22%2C%22lastName%22%3A%22Torres%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Yimeng%22%2C%22lastName%22%3A%22Zeng%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Fangping%22%2C%22lastName%22%3A%22Wan%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Natalie%22%2C%22lastName%22%3A%22Maus%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Jacob%22%2C%22lastName%22%3A%22Gardner%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Cesar%20de%20la%22%2C%22lastName%22%3A%22Fuente-Nunez%22%7D%5D%2C%22abstractNote%22%3A%22Antimicrobial%20resistance%20%28AMR%29%20poses%20a%20critical%20global%20health%20threat%2C%20underscoring%20the%20urgent%20need%20for%20innovative%20antibiotic%20discovery%20strategies.%20While%20recent%20advances%20in%20peptide%20design%20have%20yielded%20numerous%20antimicrobial%20agents%2C%20optimizing%20these%20molecules%20experimentally%20remains%20challenging%20due%20to%20unpredictable%20and%20resource-intensive%20trial-and-error%20approaches.%20Here%2C%20we%20introduce%20APEX%20Generative%20Optimization%20%28APEXGO%29%2C%20a%20generative%20artificial%20intelligence%20%28AI%29%20framework%20that%20integrates%20a%20transformer-based%20variational%20autoencoder%20with%20Bayesian%20optimization%20to%20design%20and%20optimize%20antimicrobial%20peptides.%20Unlike%20traditional%20supervised%20learning%20approaches%20that%20screen%20fixed%20databases%20of%20existing%20molecules%2C%20APEXGO%20generates%20entirely%20novel%20peptide%20sequences%20through%20arbitrary%20modifications%20of%20template%20peptides%2C%20representing%20a%20paradigm%20shift%20in%20peptide%20design%20and%20antibiotic%20discovery.%20Our%20framework%20introduces%20a%20new%20peptide%20variational%20autoencoder%20with%20design%20and%20diversity%20constraints%20to%20maintain%20similarity%20to%20specific%20templates%20while%20enabling%20sequence%20innovation.%20This%20work%20represents%20the%20first%20in%20vitro%20and%20in%20vivo%20experimental%20validation%20of%20generative%20Bayesian%20optimization%20in%20any%20setting.%20Using%20ten%20de-extinct%20peptides%20as%20templates%2C%20APEXGO%20generated%20optimized%20derivatives%20with%20enhanced%20antimicrobial%20properties.%20We%20synthesized%20100%20of%20these%20optimized%20peptides%20and%20conducted%20comprehensive%20in%20vitro%20characterizations%2C%20including%20assessments%20of%20antimicrobial%20activity%2C%20mechanism%20of%20action%2C%20secondary%20structure%2C%20and%20cytotoxicity.%20Notably%2C%20APEXGO%20achieved%20an%20outstanding%2085%25%20ground-truth%20experimental%20hit%20rate%20and%20a%2072%25%20success%20rate%20in%20enhancing%20antimicrobial%20activity%20against%20clinically%20relevant%20Gram-negative%20pathogens%2C%20outperforming%20previously%20reported%20methods%20for%20antibiotic%20discovery%20and%20optimization.%20In%20preclinical%20mouse%20models%20of%20Acinetobacter%20baumannii%20infection%2C%20several%20AI-optimized%20molecules%5Cu2014most%20notably%20derivatives%20of%20mammuthusin-3%20and%20mylodonin-2%5Cu2014exhibited%20potent%20anti-infective%20activity%20comparable%20to%20or%20exceeding%20that%20of%20polymyxin%20B%2C%20a%20widely%20used%20last-resort%20antibiotic.%20These%20findings%20highlight%20the%20potential%20of%20APEXGO%20as%20a%20novel%20generative%20AI%20approach%20for%20peptide%20design%20and%20antibiotic%20optimization%2C%20offering%20a%20powerful%20tool%20to%20accelerate%20antibiotic%20discovery%20and%20address%20the%20escalating%20challenge%20of%20AMR.%22%2C%22genre%22%3A%22%22%2C%22repository%22%3A%22bioRxiv%22%2C%22archiveID%22%3A%22%22%2C%22date%22%3A%222024-11-27%22%2C%22DOI%22%3A%2210.1101%5C%2F2024.11.27.625757%22%2C%22citationKey%22%3A%22%22%2C%22url%22%3A%22https%3A%5C%2F%5C%2Fwww.biorxiv.org%5C%2Fcontent%5C%2F10.1101%5C%2F2024.11.27.625757v1%22%2C%22language%22%3A%22en%22%2C%22collections%22%3A%5B%225EMCRXM5%22%2C%224WTWCSN6%22%5D%2C%22dateModified%22%3A%222025-02-05T14%3A22%3A43Z%22%7D%7D%2C%7B%22key%22%3A%2236BBTWWA%22%2C%22library%22%3A%7B%22id%22%3A16021941%7D%2C%22meta%22%3A%7B%22creatorSummary%22%3A%22Wenger%20et%20al.%22%2C%22parsedDate%22%3A%222024-11-01%22%2C%22numChildren%22%3A2%7D%2C%22bib%22%3A%22%3Cdiv%20class%3D%5C%22csl-bib-body%5C%22%20style%3D%5C%22line-height%3A%202%3B%20padding-left%3A%201em%3B%20text-indent%3A-1em%3B%5C%22%3E%5Cn%20%20%3Cdiv%20class%3D%5C%22csl-entry%5C%22%3EWenger%2C%20J.%2C%20Wu%2C%20K.%2C%20Hennig%2C%20P.%2C%20Gardner%2C%20J.%20R.%2C%20Pleiss%2C%20G.%2C%20%26amp%3B%20Cunningham%2C%20J.%20P.%20%282024%29.%20%3Ci%3EComputation-Aware%20Gaussian%20Processes%3A%20Model%20Selection%20And%20Linear-Time%20Inference%3C%5C%2Fi%3E%20%28No.%20arXiv%3A2411.01036%29.%20arXiv.%20%3Ca%20class%3D%27zp-DOIURL%27%20href%3D%27https%3A%5C%2F%5C%2Fdoi.org%5C%2F10.48550%5C%2FarXiv.2411.01036%27%3Ehttps%3A%5C%2F%5C%2Fdoi.org%5C%2F10.48550%5C%2FarXiv.2411.01036%3C%5C%2Fa%3E%3C%5C%2Fdiv%3E%5Cn%3C%5C%2Fdiv%3E%22%2C%22data%22%3A%7B%22itemType%22%3A%22preprint%22%2C%22title%22%3A%22Computation-Aware%20Gaussian%20Processes%3A%20Model%20Selection%20And%20Linear-Time%20Inference%22%2C%22creators%22%3A%5B%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Jonathan%22%2C%22lastName%22%3A%22Wenger%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Kaiwen%22%2C%22lastName%22%3A%22Wu%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Philipp%22%2C%22lastName%22%3A%22Hennig%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Jacob%20R.%22%2C%22lastName%22%3A%22Gardner%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Geoff%22%2C%22lastName%22%3A%22Pleiss%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22John%20P.%22%2C%22lastName%22%3A%22Cunningham%22%7D%5D%2C%22abstractNote%22%3A%22Model%20selection%20in%20Gaussian%20processes%20scales%20prohibitively%20with%20the%20size%20of%20the%20training%20dataset%2C%20both%20in%20time%20and%20memory.%20While%20many%20approximations%20exist%2C%20all%20incur%20inevitable%20approximation%20error.%20Recent%20work%20accounts%20for%20this%20error%20in%20the%20form%20of%20computational%20uncertainty%2C%20which%20enables%20--%20at%20the%20cost%20of%20quadratic%20complexity%20--%20an%20explicit%20tradeoff%20between%20computation%20and%20precision.%20Here%20we%20extend%20this%20development%20to%20model%20selection%2C%20which%20requires%20significant%20enhancements%20to%20the%20existing%20approach%2C%20including%20linear-time%20scaling%20in%20the%20size%20of%20the%20dataset.%20We%20propose%20a%20novel%20training%20loss%20for%20hyperparameter%20optimization%20and%20demonstrate%20empirically%20that%20the%20resulting%20method%20can%20outperform%20SGPR%2C%20CGGP%20and%20SVGP%2C%20state-of-the-art%20methods%20for%20GP%20model%20selection%2C%20on%20medium%20to%20large-scale%20datasets.%20Our%20experiments%20show%20that%20model%20selection%20for%20computation-aware%20GPs%20trained%20on%201.8%20million%20data%20points%20can%20be%20done%20within%20a%20few%20hours%20on%20a%20single%20GPU.%20As%20a%20result%20of%20this%20work%2C%20Gaussian%20processes%20can%20be%20trained%20on%20large-scale%20datasets%20without%20significantly%20compromising%20their%20ability%20to%20quantify%20uncertainty%20--%20a%20fundamental%20prerequisite%20for%20optimal%20decision-making.%22%2C%22genre%22%3A%22%22%2C%22repository%22%3A%22arXiv%22%2C%22archiveID%22%3A%22arXiv%3A2411.01036%22%2C%22date%22%3A%222024-11-01%22%2C%22DOI%22%3A%2210.48550%5C%2FarXiv.2411.01036%22%2C%22citationKey%22%3A%22%22%2C%22url%22%3A%22http%3A%5C%2F%5C%2Farxiv.org%5C%2Fabs%5C%2F2411.01036%22%2C%22language%22%3A%22%22%2C%22collections%22%3A%5B%225EMCRXM5%22%2C%224WTWCSN6%22%5D%2C%22dateModified%22%3A%222025-02-05T14%3A23%3A00Z%22%7D%7D%5D%7D
Torres, M. D. T., Zeng, Y., Wan, F., Maus, N., Gardner, J., & Fuente-Nunez, C. de la. (2024). A generative artificial intelligence approach for antibiotic optimization. bioRxiv. https://doi.org/10.1101/2024.11.27.625757
Wenger, J., Wu, K., Hennig, P., Gardner, J. R., Pleiss, G., & Cunningham, J. P. (2024). Computation-Aware Gaussian Processes: Model Selection And Linear-Time Inference (No. arXiv:2411.01036). arXiv. https://doi.org/10.48550/arXiv.2411.01036