Identification of Conformational Variants for Bradykinin Biomarker Peptides from a Biofluid Using a Nanopore and Machine Learning.

There is a current need to develop methods for the sensitive detection of peptide biomarkers in complex mixtures of molecules, such as biofluids, to enable early disease detection. Moreover, to our knowledge, there is currently no detection method capable of identifying the different conformations of a peptide biomarker differing by a single amino acid. Single-molecule nanopore sensing promises to provide this level of resolution. In order to be able to identify these differences in a biofluid such as serum, it is necessary to carefully characterize electrical parameters to obtain specific signatures of each biomarker population observed. We are interested here in a family of peptide biomarkers, kinins such as bradykinin and des-Arg9 bradykinin, that are involved in many disabling pathologies (allergy, asthma, angioedema, sepsis, or cancer). We show the proof of concept for direct identification of these biomarkers in serum at the single-molecule level using a protein nanopore. Each peptide exhibits two unique electrical signatures attributed to specific conformations in bulk. The same signatures are found in serum, allowing their discrimination and identification in a complex mixture such as biofluid. To extend the utility of our experimental results, we developed a principal component analysis approach to define the most relevant electrical parameters for their identification. Finally, we used semisupervised classification to assign each event type to a specific biomarker at physiological serum concentration. In the future, single-molecule scale analysis of peptide biomarkers using a powerful nanopore coupled with machine learning will facilitate the identification and quantification of other clinically relevant biomarkers from biofluids.