Related Experiment Video
Updated: Oct 10, 2026
Development and Functionalization of Electrolyte-Gated Graphene Field-Effect Transistor for Biomarker Detection
Published on: February 1, 2022
DFT and machine learning-assisted investigation of transition metal-doped γ-graphyne for DNA rare base sensing
Wenjin Miao1, Yunhan Gao1, YiHan Liu1
1Shaanxi Key Laboratory of Catalysis, School of Chemical & Environment Science, Shaanxi University of Technology, Hanzhong, 723001, China.
Background:
The development of high-performance biosensor materials for the selective detection of rare DNA bases is of great significance. Transition metal doping provides an effective strategy for regulating the electronic properties and adsorption behavior of two-dimensional carbon-based materials. However, the influence of different transition metals on the adsorption selectivity of rare DNA bases remains unclear.
Methods:
In this study, the adsorption behavior of five rare DNA bases, including cytosine (Cyt), 5-methylcytosine (5meCyt), 5-carboxylcytosine (5caCyt), 5-formylcytosine (5fCyt), and 5-hydroxymethylcytosine (5hmCyt), on transition metal-doped γ-graphyne (TM-GY) surfaces was systematically investigated using first-principles calculations. Nine transition metals (Cr, Mn, Fe, Co, Ni, Cu, Zn, Ir, and Au) were considered, generating 45 nucleobase/TM-GY configurations. The calculated adsorption energies were further employed as a small dataset for machine learning analysis, where five regression algorithms were evaluated to identify the optimal predictive model.
Results:
The machine learning models provided a complementary data-driven analysis rather than replacing density functional theory calculations or serving as a general-purpose prediction framework. The adsorption structures were optimized to elucidate the regulatory effect of transition metal doping on biosensing performance. Among the investigated systems, Cu-doped γ-graphyne exhibited relatively weak interactions with all five DNA bases, whereas Cr-doped γ-graphyne demonstrated the strongest adsorption capability toward Cyt and the weakest interaction with 5caCyt, indicating significant adsorption selectivity. Both Cu-GY and Cr-GY systems showed weak sensing responses toward 5fCyt.
Conclusions:
This study reveals the regulatory mechanism of transition metal doping on the adsorption behavior of rare DNA bases and establishes a reliable adsorption energy prediction model based on machine learning. These findings provide theoretical insights into the rational design of high-performance γ-graphyne-based biosensor materials.
