Serologic biomarker discovery for differentiating Lyme disease from diseases with similar clinical symptoms using broad profiling of antibody binding.

Introduction: Lyme disease (LD) is a tick-borne disease that is a substantial public health burden with estimated about 0.5 million new cases per year in the US and increasing incidence. Differentiating Lyme disease, especially in its early stages, from other febrile illnesses with similar clinical symptoms (look-alike diseases) represents a significant challenge due to the lack of diagnostic tools. Current diagnostic tools based on serology were not specifically developed for differential diagnosis and show limited sensitivity in early LD resulting in high false negative rates. Methods: The work presented here focuses on a broad profiling of the humoral immune response in terms of circulating antibody repertoire in patients diagnosed with LD and a number of diseases with similar clinical symptoms. A combination of antibody binding to a library of linear, diverse peptides and machine learning methods revealed a panel of biomarker proteins from the proteome of the Borrelia burgdorferi bacterium (LD causing pathogen) that can be used to differentiate between LD and other diseases. Results: A subset of the biomarkers was independently validated and demonstrated to show robust differentiating power. Importantly, the discovered biomarkers distinguish between LD patients that previously tested negative with the current test standard (false negatives) and the look-alike diseases. Discussion: These findings are important in that the discovered biomarkers can be utilized for differential diagnosis of LD. Furthermore, because the discovery approach is agnostic, the results suggest that it can also be used for biomarker discovery of other diseases.