In this project, my mentor, Jianhua Lim, and I sought to develop an objective diagnostic platform to identify differences in oral microbiome composition between saliva samples from healthy individuals and those with major depressive disorder (MDD). With over 300 million people worldwide affected by MDD, including a significant number of children and adolescents, we believe this platform has the potential to enable early detection of MDD in younger populations, leading to more effective therapeutic interventions and improved outcomes. Current methods for diagnosing MDD operate on a rule out basis, thus introducing bias into the diagnostic process. 

Recent studies have established a connection between gut microbiota and major depressive disorder, highlighting the strong correlation within the gut-oral-brain axis. Consequently, we aimed to investigate the impact of oral microbiota on neurological and systemic disorders. Given the practicality of saliva samples in point-of-care settings, we chose to focus on oral microbiota for our research. Using whole genome shotgun sequencing, we selected the top few differently expressed bacteria, Haemophilus Parainfluenzae (HP) and Parainfluenza Nigrescens (PN), to establish a validation model for subsequent diagnostic assay development. To accurately elucidate the distinct composition of HP and PN in the oral microbiome, we developed a CRISPR diagnostic assay. DNA motifs can be detected using the CRISPR-Cas12a system, which triggers the trans-cleavage of fluorescent-quencher reporters when specific bacterial sequence motifs are present. We created a calibration curve with the synthetic DNA targets in which we knew that for MDD, HP is down regulated and PN is up regulated. The calibration curve is used to back-calculate the relative composition of HP and PN from the sample to determine if HP and PN are up or down regulated compared to healthy controls. 

Building on the CRISPR detection assay, we designed guide RNAs targeting various nucleotide sequences from HP and PN. These guide RNAs were tested with synthetic long DNA targets to determine the limit of detection (LOD) for each, aiming to identify the most sensitive biomarker combination. Following the determination of the LOD, we analyzed three different healthy saliva samples to quantify the relative abundance of HP and PN in each sample as a validation model system. Due to the low concentration of HP and PN in the oral microbiome, Recombinase Polymerase Amplification (RPA) was employed to enhance sensitivity. Our goal was to further reduce the LOD for synthetic targets to enhance the sensitivity and accuracy of the detection system. Initially, we concentrated on specific targets, either HP or PN, to establish a precise LOD. For HP, the limit of detection achieved with CRISPR RPA was approximately 1 pM.

We also developed a lateral flow assay (LFA) CRISPR platform for point-of-care diagnostics, using LFA strips to determine detection limits for HP and PN targets. Next, we'll perform RPA primer screening and refine the CRISPR-RPA assay on various saliva samples to improve accuracy. Ultimately, these research endeavors aim to create an accessible diagnostic platform for adolescents with major depressive disorder utilizing the composition of the oral microbiome.