A Systems Biology and LASSO-Based Approach to Decipher the Transcriptome-Interactome Signature for Predicting...

Area of Science:

• Genomics
• Bioinformatics
• Cancer Research

Background:

• Early diagnosis of non-small cell lung cancer (NSCLC) is limited by the absence of precise molecular signatures.
• Gene expression data and biological networks offer potential for developing predictive models.

Purpose of the Study:

• To develop a highly accurate predictive model for non-small cell lung cancer (NSCLC) using gene expression data and LASSO regression.
• To identify a molecular signature for early NSCLC detection.

Main Methods:

• Utilized TCGA and GTEx data to identify differentially expressed genes (DEGs) in NSCLC.
• Constructed a biological network and identified hub genes.
• Applied LASSO logistic regression with cross-validation to a training dataset to identify 17 predictor genes.
• Validated the model on independent test datasets.

Main Results:

• The LASSO model achieved high accuracy (0.986) and AUC-ROC (0.998) on the primary test set.
• Independent validation datasets demonstrated excellent performance, with AUC-ROC values >0.99, >0.99, and 0.95.
• A web application, NSCLCpred, was developed for NSCLC prediction.

Conclusions:

• The developed LASSO model effectively predicts non-small cell lung cancer (NSCLC) with high accuracy.
• The identified gene signature holds promise for early NSCLC diagnosis.
• The NSCLCpred web application offers a valuable tool for clinical application.