A generative framework for enhanced cell-type specificity in rationally designed mRNAs DOI Creative Commons
Matvei Khoroshkin, Arsenii Zinkevich,

Elizaveta Aristova

и другие.

bioRxiv (Cold Spring Harbor Laboratory), Год журнала: 2024, Номер unknown

Опубликована: Дек. 31, 2024

Abstract mRNA delivery offers new opportunities for disease treatment by directing cells to produce therapeutic proteins. However, designing highly stable mRNAs with programmable cell type-specificity remains a challenge. To address this, we measured the regulatory activity of 60,000 5’ and 3’ untranslated regions (UTRs) across six types developed PARADE (Prediction And RAtional DEsign UTRs), generative AI framework engineer RNA tailored type-specific activity. We validated testing 15,800 de novo-designed sequences these lines identified many that demonstrated superior specificity compared existing therapeutics. PARADE-engineered UTRs also exhibited robust tissue-specific in animal models, achieving selective expression liver spleen. leveraged enhance stability, significantly increasing protein output durability vivo. These advancements translated notable increases efficacy, as PARADE-designed oncosuppressor mRNAs, namely PTEN P16, effectively reduced tumor growth patient-derived neuroglioma xenograft models orthotopic mouse models. Collectively, findings establish versatile platform safer, more precise, therapies.

Язык: Английский

Active learning of enhancers and silencers in the developing neural retina DOI Creative Commons
Ryan Z. Friedman, Avinash Ramu,

Sara Lichtarge

и другие.

Cell Systems, Год журнала: 2025, Номер unknown, С. 101163 - 101163

Опубликована: Янв. 1, 2025

Highlights•Transcription factor binding sites activate or repress depending on context•Genomic examples are insufficient to learn how context affects sites•Active learning iteratively generates informative new training data•A CNN trained with active distinguishes activating and repressing sitesSummaryDeep is a promising strategy for modeling cis-regulatory elements. However, models genomic sequences often fail explain why the same transcription can in different contexts. To address this limitation, we developed an approach train that distinguish between enhancers silencers composed of photoreceptor cone-rod homeobox (CRX). After model nearly all bound CRX from genome, coupled synthetic biology uncertainty sampling generate additional rounds data. This allowed us data multiple massively parallel reporter assays. The ability resulting discriminate identical sequence but opposite functions establishes as effective regulatory DNA. A record paper's transparent peer review process included supplemental information.Graphical abstract

Язык: Английский

Процитировано

0
Dissecting the regulatory logic of specification and differentiation during vertebrate embryogenesis DOI Creative Commons
Jialin Liu, Sebastian M. Castillo-Hair, Lucia Y. Du

и другие.

bioRxiv (Cold Spring Harbor Laboratory), Год журнала: 2024, Номер unknown

Опубликована: Авг. 27, 2024

The interplay between transcription factors and chromatin accessibility regulates cell type diversification during vertebrate embryogenesis. To systematically decipher the gene regulatory logic guiding this process, we generated a single-cell multi-omics atlas of RNA expression early zebrafish We developed deep learning model to predict based on DNA sequence found that small number underlie cell-type-specific landscapes. While Nanog is well-established in promoting pluripotency, discovered new function priming enhancer mesendodermal genes. In addition classical stepwise mode differentiation, describe instant where pluripotent cells skip intermediate fate transitions terminally differentiate. Reconstruction interactions reveals process driven by shallow network which maternally deposited regulators activate set co-regulate hundreds differentiation Notably, misexpression these sufficient ectopically their targets. This study provides rich resource for analyzing embryonic regulation differentiation.

Язык: Английский

Процитировано

4
Modelling and design of transcriptional enhancers DOI Creative Commons
Seppe De Winter, Vasileios Konstantakos, Stein Aerts

и другие.

Nature Reviews Bioengineering, Год журнала: 2025, Номер unknown

Опубликована: Фев. 28, 2025

Язык: Английский

Процитировано

0
Using machine learning to enhance and accelerate synthetic biology DOI
Kshitij Rai, Yiduo Wang, Ronan W. O’Connell

и другие.

Current Opinion in Biomedical Engineering, Год журнала: 2024, Номер 31, С. 100553 - 100553

Опубликована: Авг. 2, 2024

Язык: Английский

Процитировано

3
Uncertainty-aware genomic deep learning with knowledge distillation DOI Creative Commons
Jessica Zhou, Kaeli Rizzo, Ziqi Tang

и другие.

bioRxiv (Cold Spring Harbor Laboratory), Год журнала: 2024, Номер unknown

Опубликована: Ноя. 15, 2024

Deep neural networks (DNNs) have advanced predictive modeling for regulatory genomics, but challenges remain in ensuring the reliability of their predictions and understanding key factors behind decision making. Here we introduce DEGU (Distilling Ensembles Genomic Uncertainty-aware models), a method that integrates ensemble learning knowledge distillation to improve robustness explainability DNN predictions. distills an DNNs into single model, capturing both average ensemble's variability across them, with latter representing epistemic (or model-based) uncertainty. also includes optional auxiliary task estimate aleatoric, or data-based, uncertainty by experimental replicates. By applying various functional genomic prediction tasks, demonstrate DEGU-trained models inherit performance benefits ensembles improved generalization out-of-distribution sequences more consistent explanations cis-regulatory mechanisms through attribution analysis. Moreover, provide calibrated estimates, conformal offering coverage guarantees under minimal assumptions. Overall, paves way robust trustworthy applications deep genomics research.

Язык: Английский

Процитировано

0
A generative framework for enhanced cell-type specificity in rationally designed mRNAs DOI Creative Commons
Matvei Khoroshkin, Arsenii Zinkevich,

Elizaveta Aristova

и другие.

bioRxiv (Cold Spring Harbor Laboratory), Год журнала: 2024, Номер unknown

Опубликована: Дек. 31, 2024

Abstract mRNA delivery offers new opportunities for disease treatment by directing cells to produce therapeutic proteins. However, designing highly stable mRNAs with programmable cell type-specificity remains a challenge. To address this, we measured the regulatory activity of 60,000 5’ and 3’ untranslated regions (UTRs) across six types developed PARADE (Prediction And RAtional DEsign UTRs), generative AI framework engineer RNA tailored type-specific activity. We validated testing 15,800 de novo-designed sequences these lines identified many that demonstrated superior specificity compared existing therapeutics. PARADE-engineered UTRs also exhibited robust tissue-specific in animal models, achieving selective expression liver spleen. leveraged enhance stability, significantly increasing protein output durability vivo. These advancements translated notable increases efficacy, as PARADE-designed oncosuppressor mRNAs, namely PTEN P16, effectively reduced tumor growth patient-derived neuroglioma xenograft models orthotopic mouse models. Collectively, findings establish versatile platform safer, more precise, therapies.

Язык: Английский

Процитировано

0