Early and accurate cancer diagnosis is crucial for improving patient survival. Research has demonstrated that serum mRNA levels serve as informative biomarkers for cancer detection. In this study, we designed a weighted neural network based on DNA molecules to analyze mRNA profiles in clinical blood samples. This DNA-based molecular diagnostic platform holds the potential to provide inexpensive, non-invasive, and rapid disease screening, classification, and monitoring of disease progression.

DNA oligonucleotides All DNA oligonucleotides used in the present study were synthesized and purified by Sangon Biological Engineering Technology & Services Co. Ltd. (China). Blood RNA extraction Total RNA was extracted from human whole blood (1 ml) using the Spin Column Blood Total RNA Purification Kit according to the manufacturer’s instructions. Reverse transcription We used a reverse transcription kit (Takara) to convert RNAs into cDNA

1. Constructing the classifier model Using the expression data from the TCGA database to screen for differential genes, and combining the mRNA panel that has been validated, we constructed a linear formula for the risk score and the corresponding mRNA expression, and obtained the species and weight of each mRNA for the subsequent DNA calculation (Figure1). 2. mRNA conversion and amplification Reverse transcription + LATE-PCR was used for mRNA conversion and amplification. The procedure of LATE-PCR was first optimized and verified, which enables a linear relationship between the concentration of product and the logarithm of the initial cDNA concentration. By this means of amplification, we retained information about the concentration of gene expression embedded in the blood sample, thus allowing us to perform downstream molecular calculations. 3. Construction of a DNA-weighted neural network We have constructed a novel DNA computational platform that can assign different weights to different inputs to compute different outputs, which allows us to incorporate constructed risk scores for DNA molecular computation and diseases classifications (Figure 2). 4. Clinical diagnosis procedure Finally, we developed an operational procedure that can be used for clinical prostate cancer diagnosis (Figure 3).

We obtained a suitable diagnostic panel, then optimized the amplification method so that it could adequately retain the concentration information of mRNA in blood samples. We then validated the DNA molecular classifier to ensure that it could differentiate and process multiple inputs and draw conclusions with high accuracy and robustness. Finally, we developed an operational procedure that can be used for clinical prostate cancer diagnosis.