{"ok":true,"nota":{"id":"ba7f4cc278382d2b","slug":"un-nuevo-metodo-de-optimizacion-mejora-el-diseno-de-proteinas-que-se-p-ba7f4c","titulo":"Un nuevo método de optimización mejora el diseño de proteínas que se pegan a otras moléculas","resumen":"Investigadores probaron una variante de un algoritmo para diseñar proteínas capaces de unirse a blancos específicos. Evaluaron el método en diez objetivos usando un predictor de estructuras basado en inteligencia artificial.","fuente":"biorxiv","fuenteNombre":"bioRxiv · preprints","color":"#be2628","autor":"Ross, M., Miller, D., Wells, J.","url":"https://www.biorxiv.org/content/10.64898/2026.10.22.760942v1?rss=1","imagen":"/hoy/portada/ba7f4cc278382d2b","tema":"tecnologia","subtemas":["diseño de proteínas","inteligencia artificial","biotecnología","optimización","predicción de estructuras"],"entidades":["ESMFold2","bioRxiv"],"formato":"dato","entretenimiento":false,"pubTs":"2026-10-23T00:00:00.000Z","indexTs":"2026-10-23T12:03:18.526Z","hash":"cea7122691bb836588486b92a6eb6146617c9b0e8e51605562cf97e886dabbf3","tsa":{"time":"Oct 23 12:03:54 2026 GMT","provider":"freetsa.org","sobre":"sello"},"stats":{"views":0,"likes":0,"clicks":0,"dwell":0},"tipo":"normal","registro":"hallazgo","porQueImporta":"Diseñar proteínas que se unan a un blanco preciso sirve para desarrollar medicamentos, sensores y herramientas biotecnológicas. Para alguien de acá, el avance significa que las herramientas de diseño por computadora podrían volverse más eficientes. Es un trabajo computacional y no comprueba la eficacia de las proteínas en un organismo real.","quienFirma":"bioRxiv (preprint; investigación en diseño computacional de proteínas)","confianza":"preprint","consulta":null},"sello":{"hash":"cea7122691bb836588486b92a6eb6146617c9b0e8e51605562cf97e886dabbf3","sello":"130b26b9f4e874b2c398bc7f6e235bb57914a4ae5ca3fb574b6ddb1436e3a27c","beacon":{"drand":{"cadena":"8990e7a9aaed2ffed73dbd7092123d6f289930540d7651336225dc172e51b2ce","ronda":6577533,"aleatorio":"c4aba0dd838f5aced1ee3ce73b5a8152f120dc443db083e10a01045483b7000d","tiempo":"2026-10-23T12:03:30.000Z"},"nist":{"cadena":2,"pulso":1995961,"salida":"F41B5D0D554B9AA93D86596521336D547B043E4F52A3C280F10F60EAE511ED1743D3E56670D4E53EB431734AC5B1B8CB5DCF96D1DB26F562D058604979415129","tiempo":"2026-10-23T12:03:00.000Z"},"noAntes":"2026-10-23T12:03:30.000Z"},"verificacion":{"drand":true,"nist":true},"tsa":{"time":"Oct 23 12:03:54 2026 GMT","provider":"freetsa.org","sobre":"sello"},"original":{"titulo":"Projected Gradient Descent for Hallucination-Based Protein Binder Design","descripcion":"Gradient-based hallucination is a popular approach to protein binder design that backpropagates a loss defined on a structure predictor's output to guide optimisation in a continuous relaxation of sequence space, yet the choice of optimiser for this task has received little systematic evaluation. In this work we evaluate simplex-projected gradient descent (PGD) as an alternative to the commonly used gradient descent on softmax-reparameterised sequence logits. By updating amino-acid probabilities directly, PGD avoids softmax-induced gradient attenuation and allows amino acids suppressed early in optimisation to re-enter the design. Using an ESMFold2-based pipeline on ten targets, we find that","pubDate":"2026-10-23","url":"https://www.biorxiv.org/content/10.64898/2026.10.22.760942v1?rss=1"},"indexado":"2026-10-23T12:03:18.526Z"}}