Machine Learning Recognition of Artificial DNA Sequence with Quantum Tunneling Nanogap Junction

Area of Science:

• Nanotechnology and Materials Science
• Biotechnology and Synthetic Biology
• Computational Biology and Machine Learning

Background:

• Artificial DNA synthesis is crucial for advancing fields like biotechnology, genetics, and DNA data storage.
• Accurate and rapid electrical identification of synthetic DNA is essential for its practical applications.

Purpose of the Study:

• To investigate the electrical recognition of eight artificial DNA nucleobases (xDNA and yDNA) using quantum tunneling transport and machine learning.
• To analyze the influence of electronic coupling and molecular orbital delocalization on nucleobase recognition signals.

Main Methods:

• Embedding artificial DNA nucleobases in a solid-state nanogap junction to measure quantum tunneling transport.
• Utilizing machine learning models trained on transmission and current readout data for basecalling.
• Performing ML explainability studies and quaternary classification for nucleobase differentiation.

Main Results:

• Achieved up to 100% basecalling accuracy for xDNA and 99.80% for yDNA transmission readout data using ML.
• Demonstrated higher recognition accuracy for xDNA nucleobases compared to yDNA nucleobases.
• Successfully identified complementary, purine, and pyrimidine base pair combinations with high sensitivity and F1 scores.

Conclusions:

• The study confirms the feasibility of highly sensitive and precise electrical recognition of artificial DNA nucleobases.
• This technique has the potential to revolutionize genetic research, synthetic biology, and DNA data storage.
• ML explainability revealed normalized descriptors are more effective than transmission functions for distinguishing overlapping signals.