Integrated transcriptome analysis of new genes associated with deafness verified by machine learning and electrochemical sensors

The genetic causes of hereditary hearing loss have not been fully clarified, although more than 120 loci related to deafness have been identified. The traditional single-algorithm approach may miss the key complex gene interactions that cause hearing impairment. Conventional genetic verification techniques (such as Sanger sequencing or microarrays) are time-consuming, costly and require complex laboratory environments, which limits their application in rapid clinical transformation. Electrochemical sensors have become a promising alternative to on-site genetic testing due to their high sensitivity, portability and cost-effectiveness. This study aims to identify new genes related to deafness through comprehensive transcriptome analysis, combined with machine learning and electrochemical sensor verification, providing new targets for the diagnosis and treatment of hereditary hearing loss. Gene expression data were obtained from the GEO database, including 43 samples, which were divided into three groups: deafness patients, healthy controls, and heterozygous carriers. After quality control and preprocessing of the data, multiple machine learning algorithms (including random forest, LASSO regression, XGBoost, etc.) were used for analysis, and key genes were identified through integrated methods. The application of electrochemical sensors provides a rapid, sensitive and cost-effective method for gene verification. The combination with machine learning algorithms further enhances the accuracy and reliability of gene detection. The research successfully identified new candidate genes that may be involved in the pathogenesis of deafness through a multi-algorithm bioinformatics framework. These genes are ideal targets for subsequent verification using electrochemical sensors. This comprehensive approach combines computational prediction with scalable clinical verification, paving the way for the use of electrochemical sensors for gene detection in resource-limited environments. It is expected to promote the early diagnosis and treatment of hereditary hearing loss. • Electrochemical biosensors have emerged as a promising alternative due to their high sensitivity, portability, and cost-effectiveness for point-of-care gene detection. • Machine learning models achieved robust performance • Our multi-algorithm bioinformatics framework successfully identified novel candidate genes potentially involved in deafness pathogenesis.