Exploring microRNA-mediated alteration of EGFR signaling pathway in non-small cell lung cancer using an mRNA:miRNA regression model supported by target prediction databases

• We construct mRNA:miRNA regression models exploring mRNA and miRNA relationship.
• R2 and LOOCV demonstrate good performances of these models.
• The model is suitable for predicting the expression level of mRNA using miRNA(s).
• The model is able to identify miRNA roles as tumor suppressor mirs or oncomirs.
• The mode can also be used to explore the underlying mechanisms of NSCLC.

EGFR signaling pathway and microRNAs (miRNAs) are two important factors for development and treatment in non-small cell lung cancer (NSCLC). Microarray analysis enables the genome-wide expression profiling. However, the information from microarray data may not be fully deciphered through the existing approaches. In this study we present an mRNA:miRNA stepwise regression model supported by miRNA target prediction databases. This model is applied to explore the roles of miRNAs in the EGFR signaling pathway. The results show that miR-145 is positively associated with epidermal growth factor (EGF) in the pre-surgery NSCLC group and miR-199a-5p is positively associated with EGF in the post-surgery NSCLC group. Surprisingly, miR-495 is positively associated with protein tyrosine kinase 2 (PTK2) in both groups. The coefficient of determination (R2) and leave-one-out cross-validation (LOOCV) demonstrate good performance of our regression model, indicating that it can identify the miRNA roles as oncomirs and tumor suppressor mirs in NSCLC.