MARLOWE: Taxonomic Characterization of Unknown Samples for Forensics Using De Novo Peptide Identification

The preprint operates within the paradigm of forensic proteomics and metaproteomics, specifically focusing on taxonomic characterization of unknown biological samples using mass spectrometry data.

This preprint presents MARLOWE, a new computational tool for taxonomic characterization using de novo peptide sequencing combined with a statistical approach. While it introduces a novel method within the existing paradigm, it doesn't challenge the fundamental assumptions of the field. The research addresses a current gap in forensic proteomics by attempting to use de novo sequencing coupled with tag extraction, peptide strength assessment and a statistical approach to reduce reliance on traditional database searching methods, thereby optimizing specificity and characterization capabilities. The authors acknowledge the current limitations of de novo peptide sequencing, which are reflected in the lower accuracy compared to database search methods like MiCId. However, the high specificity achieved by MARLOWE, especially within the B. cereus group, signifies a valuable contribution to the field as it focuses on more accurate organism identification. The preprint explores performance within various bacterial groups, including closely related species and addresses the challenges associated with shared peptides and database limitations. This work is characteristic of normal science, refining and extending existing tools and techniques to improve accuracy and address specific challenges. It builds upon established methods and aims to enhance current practices within the field. Therefore, I've classified it as Normal Science with high certainty.