Background/Objectives: Predicting the effects of protein and DNA mutations on the binding free energy of protein–DNA complexes is crucial for understanding how DNA variants impact wild-type cellular function. As many cellular interactions involve protein–DNA binding, accurately predicting changes in binding free energy (ΔΔG) is valuable for distinguishing pathogenic mutations from benign ones. Methods: This study describes the development and optimization of the SAMPDI-3Dv2 machine learning method, which is trained on an expanded database of experimentally measured ΔΔGs. This enhanced model incorporates new features, including the 3D structure of the mutant protein, features of the mutant structure, and a position-specific scoring matrix (PSSM). Benchmarking was conducted using 5-fold cross-validation. Results: The updated SAMPDI-3D model (SAMPDI-3Dv2) achieved Pearson correlation coefficients (PCCs) of 0.68 for protein and 0.80 for DNA mutations. These results represent significant improvements over existing tools. Additionally, the method’s rapid execution time enables genome-scale predictions. Conclusions: The improved SAMPDI-3Dv2 shows enhanced predictive performance for analyzing mutations in protein–DNA complexes. By leveraging structural information and an expanded training dataset, SAMPDI-3Dv2 provides researchers with a more accurate and efficient tool for mutation analysis, contributing to identifying pathogenic variants and improving our understanding of cellular function.

Calpain-1 and calpain-2 are heterodimeric proteases consisting of a common small regulatory subunit CAPNS1 and a large catalytic subunit, CAPN1 or CAPN2, respectively. These calpains have emerged as potential therapeutic targets in cancer and other diseases through their roles in cell signaling pathways affecting sensitivity to chemotherapeutic and targeted drugs, and in promoting metastasis. While inhibition of calpains has the potential to provide therapeutic benefit to cancer patients, there are currently no clinically approved active site directed drugs that specifically and effectively inhibit them. However, the structures of calpain-1 and calpain-2 make them susceptible to allosteric inhibition aimed at interfering with heterodimerization of the catalytic and regulatory subunits, which is necessary for stability and proteolytic activity. Split-Nanoluciferase biosensors were generated to quantify the protein-protein interactions (PPIs) between the calcium-binding penta-EF hand (PEF) domains of CAPN1 or CAPN2 and CAPNS1. These biosensors were used to quantify the heterodimer dissociation constants (K_D) of calpain-1 and calpain-2, estimated at 185 nM and 509 nM, respectively, in the presence of 5 mM Ca²⁺; and 362 nM and 1651 nM, respectively, in the presence of Mg²⁺. The half-maximal Ca²⁺ concentrations supporting these PPIs for calpain-1 and calpain-2 were 59.9 μM and 940.8 μM, respectively. Molecular modeling, based on the crystal structure of calpain-2, was used to predict 20 residues of the PEF domains that contribute to heterodimerization. Individual point mutation of CAPNS1 at Q263 reduced the catalytic activity of calpain-2 to 51.0 ± 6.4 % in live cells.

Background/Objectives: Predicting the effects of mutations in proteins and DNA on the binding free energy of protein-DNA complexes is crucial for understanding how DNA variants impact wild-type cellular function. As many cellular interactions involve protein-DNA binding, accurately predicting changes in binding free energy (ΔΔG) is valuable for distinguishing pathogenic mutations from benign ones. Methods: The study describes the further development of the SAMPDI-3D machine learning method, which is trained on an expanded database of experimentally measured ΔΔGs. This enhanced model incorporates new features, including the 3D structure of the mutant protein, structural features of the mutant structure, and a position-specific scoring matrix (PSSM). Benchmarking was conducted using 5-fold cross-validation. Results: The updated SAMPDI-3D model achieves a Pearson Correlation Coefficient (PCC) of 0.68 for mutations in proteins and a PCC of 0.80 for mutations in DNA. These results represent significant improvements over existing tools. Additionally, the method’s rapid execution time enables genome-scale predictions. Conclusions: The advancements in SAMPDI-3D enhance its predictive performance and utility for analyzing mutations in protein-DNA complexes. By leveraging structural information and an expanded training dataset, SAMPDI-3D provides researchers with a more accurate and efficient tool for mutation analysis, contributing to the identification of pathogenic variants and improving our understanding of cellular function.