Background Hundreds of SCN5A nonsynonymous single nucleotide variants (nsSNVs) have been identified in long QT syndrome (LQTS) and Brugada syndrome (BrS) cases. However, a 2% background rate of rare SCN5A nsSNVs among healthy white controls results in a signal-to-noise ratio (SNR) of 5:1 for LQT3 and 10:1 for BrS1, thus confounding clinical genetic testing. Here, we determine whether a set of 7 in silico prediction tools can enhance the SNR associated with LQTS/BrS genetic testing. Methods In this study, (1) conservation across species, (2) paralogs, (3) Grantham matrix values, (4) SIFT, (5) PolyPhen2, (6) Consensus Deleteriousness score (Condel), and (7) Mutation Assessor (Mass) algorithms were used to assign pathogenic or benign status to nsSNVs identified across 388 clinically definite LQTS index cases, 2500 suspected LQTS cases, 2111 BrS cases, the 1000 Genome project (n = 1092), 1380 ostensibly healthy control subjects, and the NHLBI Exome Sequencing Project (n = 6502). The estimated predictive values (EPVs) were determined for each tool independently, in concert with previously published protein topology-derived EPVs, and synergistically when >3 tools were in agreement. Results Although all 7 tools displayed a statistically significant ability to distinguish between SCN5A nsSNVs identified in cases from those in controls, a combination of >3 tools in agreement resulted in the greatest polarization of EPVs [>3 EPVs: LQTS=87 (81–90), BrS=96 (94–97); ≤3 EPVs: LQT=0 (0–35), BrS=21 (0–52)]. For LQTS and BrS, the in silico tools were able to enhance mutation calling within the interdomain linkers [>3 EPVs: LQTS=85 (70–93), BrS=86 (71–94)]. Interestingly for LQTS, transmembrane domains I and III had low EPVs [64 (27–82) and 68 (28-86), respectively]. The addition of in silico tools nearly doubles the SNR for both LQTS and BrS. Conclusions Although individual in silico tools alone can help upgrade/downgrade the pathogenicity of SCN5A nsSNVs, the development of tools that couple multiple independent algorithms appears more promising as the synergistic use of multiple existing in silico tools enhances the classification of nsSNVs that reside within regions where the topology-based probability of pathogenicity is suboptimal.