Forensic DNA analysis is well established for phenotyping, providing valuable investigative leads. Proteomics, the large-scale study of proteins, is emerging as a complementary tool to DNA analysis, particularly for enhancing the evidential value of traces. This study explores the potential of proteomics to extract phenotypic traits from whole blood, using the estimation of biological sex as a starting point. Using LC–MS/MS, proteomes from 100 whole blood samples of known sex were analyzed to develop a biological sex classifier. Cross-validation of the model highlighted the model’s ability to achieve accurate classification, identifying key peptides, such as those from pregnancy zone protein and ceruloplasmin, as critical markers. To test real-world applicability, mock case samples were generated, bringing attention to the need for model robustness. Overall, our results suggest that using proteomics to infer phenotypic traits from whole blood samples in the context of forensics, while feasible, is hindered by hard-to-overcome technical challenges. We therefore recommend that future forensic proteomics research be directed toward areas where it can be most informative, such as source attribution and estimating the timeline of events, rather than focusing on extracting phenotypic traits for donor profiling.