Design and Optimization of Ultrasound-Assisted Nucleic Acid Extraction System

Nucleic acid extraction is the first step in nucleic acid analysis. Currently, designing an efficient, robust, and automated nucleic acid extraction system has been a challenging topic. Moreover, nucleic acid extraction system is commonly designed based on the magnetic beads method which is the most widely used in clinical practice, yet there is a lack of methods to characterize and measure the beads and fluid motion processes. This leads researchers to rely on biological experimental results for evaluating the performance of the system. To address these issues, we proposed a highly efficient and automated nucleic acid extraction system, including an ultrasound-assisted extraction instrument and a microfluidic chip. This instrument integrated ultrasound technology throughout the nucleic acid extraction process, significantly enhanced efficiency while eliminating the need for complex functional modules or additional reagent materials. It can complete extraction within 12 minutes, whereas manual methods or other instruments typically take 30 minutes and more. Furthermore, we presented a non-contact flow pattern measurement approach that combines acoustic knowledge and machine learning, which can be used to capture fluid motion within a miniature confined chamber under ultrasonic stimulation. Multidimensional experiments were conducted on the system. Firstly, we annotated 1946 flow field images and achieved an average recognition precision of 96.4%. Based on measurements and analysis of the actual flow fields, enhancements were made to the ultrasound stimulation strategy, which significantly improved system performance. As a final point, the biological experimental verification results indicated that the system performed comparably to the "gold standard" manual nucleic acid extraction method. The Coefficient of Variation (CV) remained under the 5%, signaling excellent stability. Particularly, it demonstrated a 100% detection rate for low-concentration samples (1.7 × 10 ³ copies/mL).