Optimization of bioinformatics parameters for DIA-based proteomics in targeted protein degradation

Targeted Protein Degradation (TPD) is a transformative strategy for targeting oncoproteins previously considered undruggable. A central challenge in translating TPD compounds into viable cancer therapeutics is the accurate delineation of on-target efficacy from off-target toxicities, which is critical for patient safety and therapeutic efficacy. Data-Independent Acquisition (DIA) mass spectrometry provides the profound sensitivity and comprehensive proteome coverage necessary for systematically mapping protein abundance changes induced by TPD compounds. However, the accurate assessment of on-target efficacy and off-target effects critically depends on the selection of appropriate bioinformatic parameters in DIA proteomics data analysis. To address this, we conducted a systematic evaluation of key bioinformatic parameters. Our findings identify Unique Peptide filtering and Imputation as the most influential factors in precisely defining both on-target efficacy and off-target effects. Specifically, applying a UniquePep threshold ≥2 effectively minimized off-target misidentification. For imputation methods: when data was completely missing in treatment groups, row_min imputation yielded non-significant p-values; when data was partially missing, min imputation resulted in high intra-group variance with non-significant p-values. Therefore, we recommend row_min_half as a general-purpose imputation approach. Furthermore, we investigated the impact of peptide-level bioinformatic parameters on on-target efficacy and off-target effects. Our analysis revealed that high-abundance outlier peptides in control groups, low peptide detection rates in samples, and the selection of specific peptides to represent protein abundance significantly influence result accuracy. These confounding factors can be mitigated by filtering high-abundance outlier peptides, improving peptide detection rates, and implementing appropriate peptide-level missing value imputation, thereby optimizing the reliability of TPD outcomes. This study provides a critical, optimized framework for DIA proteomics analysis specifically tailored to TPD research. By ensuring accurate assessment of degrader specificity, this pipeline will accelerate the prioritization of lead compounds and de-risk the development of safer, more effective targeted protein degraders for cancer therapy.