The directed evolution of biomolecules is an iterative process. Although advancements in language models have expedited protein evolution, effectively evolving RNA remains a challenge.
The directed evolution of biomolecules is an iterative process. Although advancements in language models have expedited protein evolution, effectively evolving RNA remains a challenge. RNA aptamers, selected for their binding properties, provide an ideal system to address this challenge, yet traditional aptamer discovery still relies on labor-intensive, multi-round screening.
Here we introduce GRAPE-LM , a generative artificial intelligence framework designed for the one-round evolution of RNA aptamers. GRAPE-LM integrates a transformer-based conditional autoencoder with nucleic acid language models and is guided by CRISPR−Cas-based aptamer screening data derived from intracellular environments. We validate GRAPE-LM on three disparate targets: the human T cell receptor CD3ε, the receptor-binding domain of the SARS-CoV-2 spike protein and the human oncogenic transcription factor c-Myc . GRAPE-LM, informed with only a single round of CRISPR−Cas-based screening, successfully obtains RNA aptamers that outperform those driven from multiple rounds of human selection and optimization.Fig. 1: Illustration of GRAPE-LM and computational results.Fig. 3: Benchmarking GRAPE-LM-derived aptamer leads targeting CD3ε.https://grape-lm.bioailab.net/Google ScholarGoogle ScholarLee, G., Jang, G. H., Kang, H. Y. & Song, G. Predicting aptamer sequences that interact with target proteins using an aptamer-protein interaction classifier and a Monte Carlo tree search approach.Alipanahi, B., Delong, A., Weirauch, M. T. & Frey, B. J. Predicting the sequence specificities of DNA- and RNA-binding proteins by deep learning.Nithin, C., Kmiecik, S., Błaszczyk, R., Nowicka, J. & Tuszyńska, I. Comparative analysis of RNA 3D structure prediction methods: towards enhanced modeling of RNA–ligand interactions.Torres, M. D. T., Chen, L. T., Wan, F., Chatterjee, P. & de la Fuente-Nunez, C. Generative latent diffusion language modeling yields anti-infective synthetic peptides.Patel, S., Peng, F. Z., Fraser, K., Chatterjee, P. & Yao, S. EvoFlow-RNA: generating and representing non-coding RNA with a language model. Preprint at Penić, R. J., Vlašić, T., Huber, R. G., Wan, Y. & Šikić, M. RiNALMo: general-purpose RNA language models can generalize well on structure prediction tasks.Ishida, R. et al. RaptRanker: in silico RNA aptamer selection from HT-SELEX experiment based on local sequence and structure information.Li, W. & Godzik, A. Cd-hit: a fast program for clustering and comparing large sets of protein or nucleotide sequences.Freage, L., Jamal, D., Williams, N. B. & Mallikaratchy, P. R. A homodimeric aptamer variant generated from ligand-guided selection activates the T cell receptor cluster of differentiation 3 complex.Zumrut, H. E. et al. Integrating ligand-receptor interactions and in vitro evolution for streamlined discovery of artificial nucleic acid ligands.Nakhjavani, M. et al. A flow cytometry-based cell surface protein binding assay for assessing selectivity and specificity of an anticancer aptamer.Li, J., Zhang, S., Zhang, D. & Chen, S.-J. Vfold-Pipeline: a web server for RNA 3D structure prediction from sequences.Li, J. et al. Diverse high-affinity DNA aptamers for wild-type and B.1.1.7 SARS-CoV-2 spike proteins from a pre-structured DNA library.Yang, G. et al. Identification of SARS-CoV-2-against aptamer with high neutralization activity by blocking the RBD domain of spike protein 1.Alves Ferreira-Bravo, I. & DeStefano, J. J. Xeno-nucleic acid 2′-fluoro-arabino nucleic acid aptamers to the receptor-binding domain of SARS-CoV-2 S protein block ACE2 binding.Saify Nabiabad, H., Amini, M. & Demirdas, S. Specific delivering of RNAi using spike’s aptamer-functionalized lipid nanoparticles for targeting SARS-CoV-2: a strong anti-Covid drug in a clinical case study.Dieng, A. B., Kim, Y., Rush, A. M. & Blei, D. M. Avoiding latent variable collapse with generative skip models. In 2397–2405 .Zhang, C., Shine, M., Pyle, A. M. & Zhang, Y. US-align: universal structure alignments of proteins, nucleic acids, and macromolecular complexes.Ni, B., Kaplan, D. L. & Buehler, M. J. ForceGen: end-to-end de novo protein generation based on nonlinear mechanical unfolding responses using a language diffusion model.Nguyen Quang, N., Bouvier, C., Henriques, A., Lelandais, B. & Ducongé, F. Time-lapse imaging of molecular evolution by high-throughput sequencing.Mathews, D. H., Sabina, J., Zuker, M. & Turner, D. H. Expanded sequence dependence of thermodynamic parameters improves prediction of RNA secondary structure.Molejon, N. A. et al. Selection of G-rich ssDNA aptamers for the detection of enterotoxins of the cholera toxin family.We thank all colleagues in our laboratories for experimental assistance and helpful discussions. This work was supported by the National Key Research and Development Program of China , the National Natural Science Foundation of China , the Department of Science and Technology of Guangdong Province , the Guangdong Basic and Applied Basic Research Foundation , the Shenzhen Science and Technology Program , the Shenzhen Stable Support Grant , the Shenzhen Science and Technology Program , the China Postdoctoral Science Foundation , the Postdoctoral Fellowship Program of the China Postdoctoral Science Foundation and the Internal Fund of the National Engineering Laboratory for Big Data System Computing Technology .State Key Laboratory of Quantitative Synthetic Biology, Shenzhen Institute of Synthetic Biology, Shenzhen Institute of Advanced Technology, Chinese Academy of Sciences, Shenzhen, China Y.W., Y.Z. and Jun Zhang conceived and designed the study. Jun Zhang, S.T., H.Z. and X.M. performed modeling and computational analyses. Ju Zhang, C.L., Y.C. and B.L. performed wet lab experiments. All authors analyzed the data and interpreted the results. Y.W., Y.Z., Jun Zhang and Ju Zhang wrote the manuscript, with contributions from all coauthors. Jun Zhang and Ju Zhang contributed equally to this work. Y.W. and Y.Z. supervised the project. All authors read and approved the final manuscript.) Comparison of activity-guided versus sequence similarity-based semantic spaces. The activity-guided latent space enables focused sampling of functional aptamers, while traditional sequence similarity-based spaces permit only random exploration. Data processing pipeline. Training and test datasets were derived from the first-round output of CRISmers . Unique sequences were assigned pseudo-activity scores based on enrichment frequencies.is a hyperparameter that needs to be optimized for different targets, and it is recommended to initially use 0.05. The sequences were clustered using CD-HIT with a threshold of 0.8. Sequences only with a high confidence were used . For targets with fewer initial sequences, the threshold of individual unique sequences was set to 1. The final selected sequences were then randomly divided into training and test groups in an 8:2 ratio to construct the training and test sets. Icons in figures were created with BioRender . Data plot mean with standard deviation , n = 3 experimental replicates. The statistical analyses were conducted using a two-sided Student’s t-test .) Dose-response binding curves of the Cy5-labeled CD3ε aptamers on Jurkat cells. CD3ε knock-out cells serve as negative control to assess binding specificity. Apparent equilibrium dissociation constants were derived from nonlinear regression analysis using GraphPad Prism. Data represent mean ± SD .) Luciferase reporter assay results of the top 50 candidate sequences from one round of CRISmers screening, ranked by sequencing read abundance. Data represent mean with SD, n = 3 biological replicates. Comparative GFP reporter assay of libraries from one-round screening using CRISmers versus GRAPE-LM. “Mock” denotes transfection with unrelated control plasmids; “CRISmers” represents transfection with the sub-library constructed after one round of CRISmers screening; “GRAPE-LM” corresponds to transfection with the library generated by GRAPE-LM in a single round.) MST binding curves of previously reported SELEX-derived aptamers targeting the RBD of SARS-CoV-2 Spike protein. Data showed mean with SD, with n = 3 biological replicates. Results of Kd value determination based on MST for the Lead of 2nd round and the Lead of 5th round from iterative CRISmers. Data showed mean with SD, with n = 3 biological replicates. The receptor-binding domain resides in the S1 subunit of the SARS-CoV-2 Spike glycoprotein. To validate binding specificity during MST assays, the S2 subunit—a non-target structural region—was used as a negative control, ensuring the aptamer selectively recognizes the RBD rather than unrelated epitopes.Extended Data Fig. 7 Case study of the internal loop motif in two representative aptamers.) aptamers. The dashed box highlights the computationally predicted binding site within the internal loop. Data represent the mean with SD .This new paradigm is enabled by one shot GRAPE-LM introduced in this work, informed by one round CRISmers. Icons in figures were created with BioRender holds exclusive rights to this article under a publishing agreement with the author or other rightsholder; author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
United States Latest News, United States Headlines
Similar News:You can also read news stories similar to this one that we have collected from other news sources.
New single 1.5mm sensor gives robots 180-degree vision and bionic smell powerScientists pack 1,027 lenses and a bionic nose into a 180-degree vision sensor, allowing next-gen robots to navigate and detect threats.
Read more »
AMC's The Vampire Lestat Releases First Ever Rock SingleJennifer has been a journalist for over a decade. She has covered a wide range of topics such as TV, comics, gaming, politics, and finance.
Read more »
'The Vampire Lestat' Is a Blood-Sucking Rock Star With New David Bowie-Inspired SingleSam Reid throwing his head back with a mic on-stage in Interview With a Vampire Season 3
Read more »
Eurozone: Leaders push single market planING’s Carsten Brzeski and Bert Colijn describe how EU leaders discussed restoring European competitiveness and long‑term growth at an informal summit in Belgium, which took place on February 12.
Read more »
RNA strand that can almost self-replicate may be key to life's originsLife may have begun when RNA molecules began to replicate themselves, and now we’ve finally found an RNA molecule that is very close to being able to do this
Read more »
I’m Happy Being Single, But Could I Be Singles-Night Single?Now that having a boyfriend is embarrassing and being alone is cool, is there any reason to go to a singles night?
Read more »