Accurate Estimation of Context-Dependent False Discovery Rates in Top-Down Proteomics

Area of Science:

• Proteomics
• Biochemistry
• Computational Biology

Background:

• Top-down proteomics enables high-throughput identification of thousands of proteins and proteoforms.
• Accurate false discovery rate (FDR) estimation is critical for the reliable adoption and consistency of top-down proteomics.
• Current FDR calculation methods are context-dependent, influenced by MS data quality, databases, search engines, and parameters.

Purpose of the Study:

• To design, implement, and validate a context-dependent framework for accurate FDR calculation at multiple molecular levels in top-down proteomics.
• To address the limitations of existing FDR estimation methods that may hinder the widespread use of top-down proteomics.
• To provide a scalable and post-search applicable solution for enhancing the quality of top-down proteomics results.

Main Methods:

• Development of a context-dependent FDR calculation framework tailored for top-down proteomics.
• Validation of the framework using a manually curated training set of 546 confirmed proteoforms.
• Examination of various search contexts to assess FDR calculation accuracy at different molecular levels (PrSM, protein, isoform, proteoform).

Main Results:

• A significant under-reporting of the true FDR at the protein level was observed when calculated at the PrSM level (average 24-fold difference).
• The developed framework demonstrated accurate FDR estimation across multiple molecular levels.
• The TDCD_FDR_Calculator, an open-source tool, was created to implement the context-dependent FDR calculation.

Conclusions:

• A context-dependent FDR calculation framework is essential for accurate results in top-down proteomics.
• The TDCD_FDR_Calculator provides a valuable, open-source tool for improving the reliability of proteoform identification.
• Implementing this tool can enhance the consistency and adoption of top-down proteomics across various research settings.