Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
To overcome severe performance degradation in tandem mass spectrum (MS/MS) prediction under chemical and acquisition distribution shifts, the authors present SPARC, a test-time specialization framework that adapts pretrained models using a reference spectral library without requiring ground-truth query spectra. For each input compound, the system retrieves structurally related reference spectra to recalibrate fragment intensities within latent fragmentation space, pairing this with reliability-aware consistency regularization to maintain stable representations during continual adaptation. Across MassSpecGym, NPLIB1, and GNPS benchmarks, this retrieval-guided test-time training consistently outperforms static pretrained predictors across challenging out-of-distribution transfer regimes.
Pretrained scientific foundation models can overcome massive chemical and instrumentation shifts entirely at test-time by using reference-guided retrieval to recalibrate latent predictions on the fly without query supervision.
Tandem mass spectrum prediction supports compound identification across metabolomics, natural-product discovery, and environmental analysis. However, pretrained predictors often degrade under shifts in chemical space and acquisition conditions, while retraining domain-specific models from scratch is costly. We introduce SPARC, a retrieval-guided test-time specialization framework that adapts a pretrained predictor using a spectral reference library without accessing test-query spectra. For each target query, SPARC retrieves chemically related reference spectra to recalibrate fragment intensities within the learned fragmentation space. During Transfer, SPARC combines reference-guided spectral adaptation with reliability-aware consistency, using reconstruction behavior on retrieved spectra to selectively preserve trustworthy predictions during continual specialization. Across MassSpecGym, NPLIB1 and application-specific GNPS libraries, SPARC improves spectral prediction under multiple transfer settings. These results establish retrieval-guided test-time specialization as a practical strategy for extending pretrained MS/MS predictors to specific chemical and acquisition domains, with continual test-time training providing further refinement during deployment.