LP-SDA

LP-SDA (Label Propagation-based Signal Detection Algorithm) is a software tool that enhances drug safety signal detection by combining pre-clinical drug chemical structures with post-market spontaneous reports from the FDA Adverse Event Reporting System (FAERS). The tool first computes original drug safety signals using standard signal detection algorithms and then constructs a drug similarity network based on chemical structures.

It generates enhanced drug safety signals by propagating the original signals on the drug similarity network. LP-SDA enriches post-market safety reports with pre-clinical drug similarity information, effectively addressing the issue of insufficient case reports for newly approved drugs. The tool can be applied to popular signal detection algorithms such as PRR, ROR, MGPS, and BCPNN, resulting in more accurate drug safety signals than the corresponding baselines. Additionally, LP-SDA can identify potential adverse drug reactions (ADRs) for newly approved drugs, enabling early detection of ADRs.