Environment-adaptive molecular neural networks enabled by heterogeneous threshold DNA neuron

Molecular neural networks enable autonomous information processing in biochemical environments, yet most existing DNA-based implementations operate with fixed connectivity after assembly, limiting their ability to adapt network topology to changing environments. Here, we report an environment-adaptive molecular neural network enabled by DNA neurons with sequence-programmed, heterogeneous activation thresholds. These neurons incorporate triplex-based proton-gated DNA switches with tunable pH-response windows, which regulate the accessibility of toehold and recognition domains to switch pre-encoded weighted connections between inactive and signal-propagating states, thereby reconfiguring network connectivity without changing molecular composition. We demonstrate this principle in a DNA neural network that processes 64-bit inputs, where environmental acidification activates dormant neurons and reconfigures the computational topology from a sparse convolutional neural network for coarse two-class categorization into a fully connected neural network for fine-grained discrimination among four distinct patterns. Our results establish DNA-based neural circuits capable of adapting their computational architecture to environmental cues, offering a framework for molecular information processing in fluctuating biochemical environments.

Authors

Institutions

Publication Details

Journal
Science Advances
Published
2026-11-18
DOI
https://doi.org/10.1126/sciadv.aek3182
Primary Topic
DNA and Biological Computing
Type
article
Field-Weighted Citation Impact
0.00

Funders

Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Environment-adaptive molecular neural networks enabled by heterogeneous threshold DNA neuron

Xiewei Xiong, Tong Zhu, Ziyi Huang, Hao Pei et al.
3 links
Science Advances
DNA and Biological Computing
article

Environment-adaptive molecular neural networks enabled by heterogeneous threshold DNA neuron

Xiewei Xiong, Tong Zhu, Ziyi Huang, Hao Pei, Li Li, Jiayi Chen, Wanhui Xu, Na Li, Sichen Xing
article en

Abstract

Molecular neural networks enable autonomous information processing in biochemical environments, yet most existing DNA-based implementations operate with fixed connectivity after assembly, limiting their ability to adapt network topology to changing environments. Here, we report an environment-adaptive molecular neural network enabled by DNA neurons with sequence-programmed, heterogeneous activation thresholds. These neurons incorporate triplex-based proton-gated DNA switches with tunable pH-response windows, which regulate the accessibility of toehold and recognition domains to switch pre-encoded weighted connections between inactive and signal-propagating states, thereby reconfiguring network connectivity without changing molecular composition. We demonstrate this principle in a DNA neural network that processes 64-bit inputs, where environmental acidification activates dormant neurons and reconfigures the computational topology from a sparse convolutional neural network for coarse two-class categorization into a fully connected neural network for fine-grained discrimination among four distinct patterns. Our results establish DNA-based neural circuits capable of adapting their computational architecture to environmental cues, offering a framework for molecular information processing in fluctuating biochemical environments.

Science AdvancesVol. 12(47)
Shanghai Innovation Institute (CN), East China Normal University (CN)
National Natural Science Foundation of China
DNA and Biological Computing
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon