lncAPNet enables the deciphering of lncRNA-mRNA connections in patient transcriptomic data.

Motivation: Long non-coding RNAs regulate gene expression through chromatin remodeling, transcriptional control, and post-transcriptional modulation, influencing physiological cell homeostasis but also disease onset. Yet most transcriptomic and network-based studies rely on descriptive linear co-expression analyses, missing nonlinear and mechanistic insights. Emerging ML/DL methods offer promise but remain limited by data sparsity, noise, insufficient biological priors, and poor interpretability, constraining systems-level lncRNA-mRNA motif discovery. Results: In this manuscript, we introduce lncAPNet, an extended version of the APNet workflow, which integrates graph-based nonlinear inference of lncRNA-mRNA interactions using NetBID2's and scMINERs activity logic within a lncRNA-focused SJARACNe co-expression network, coupled with PASNet, a biologically informed sparse deep learning model. This framework enables explainable identification of lncRNA drivers in three different cancer type case studies, two with bulk RNA-seq datasets [Chronic Lymphocytic Leukemia and Prostate Adenocarcinoma] and one by combining bulk RNA-seq and scRNA-seq omics datasets [Breast Invasive Carcinoma], uncovering lncRNA drivers that illuminate lncRNA-mediated programs in cancer progression. Availability and implementation: lncAPNet's R scripts, Python scripts, and Nextflow pipeline are available at the GitHub repository: https://github.com/BiodataAnalysisGroup/lncAPNet.