Preparation of sequencing libraries for multi-analyte analysis of RNAs

Abstract Circulating cell-free DNA (cfDNA) are commonly considered promising multi-cancer detection (MCD) biomarkers. The recently proposed Vanguard Study on MCD should help elucidate the effectiveness and practicality of such tests going forward. Meanwhile, there is growing evidence that cell-free RNA fragmentomes (RFs), which include small RNAs and RNA fragments, may become a novel class of liquid biopsy biomarkers providing increased sensitivity due to RNA fragmentome substantially higher abundancy than cfDNA fragmentome. Also, RFs provide information about biological function and tissues of origin due to the process of their biogenesis. The sequencing analysis of diverse pools of RFs could be integrated in multiomic approaches that would further enhance their diagnostic capabilities for cancers and other diverse pathologies. Conventional kits for sequencing library preparation are limited to the capture of RFs with 5’-p and 3’-OH ends representing only ~10% of the whole RNA fragmentome. Although several reported methods can detect all or some hidden RF species, none of them can detect all and distinguish different RF types. Also, the analysis of RFs by different methods is affected by method- specific biases. Besides the detection of all RF types, it is essential to enable analysis of different RNA classes represented by different RF types and find the type(s) that allows detection of cancer with higher sensitivity and specificity. Currently there are no commercially available kits for sequencing-based, comprehensive RF analysis. To address these technical shortcomings, we propose a novel approach, RealSeq®-RF, that enables profiling of both the total RF content and the individual RF types in RNA samples using a core library preparation kit with upfront enzymatic pretreatment(s) effecting the RNA end types. In Phase I, we developed a RealSeq®-RF prototype and demonstrated its superior ability to: 1) detect all RFs simultaneously and specifically for individual RF types; and 2) distinguish between plasma RNA samples from patients with breast cancer and healthy donors based on analysis of differently expressed type-specific RF content in comparison to benchmark Phospho-Seq method. In Phase II of this project, we plan: 1) optimize RealSeq-RF’s specificity and extend its capabilities of detecting RFs and different types and lengths in human plasma; 2) convert the optimized of RealSeq-RF into a commercially viable protocol and first-in-class kit for research and diagnostic applications; 3) adapt the protocol for automation with liquid and bead handling workstations to minimize technical variability during isolation of plasma cfRNA, pretreatments and the sequencing library preparations; 4) develop a robust and user-friendly bioinformatic pipeline to analyze RealSeq-RF sequencing data and generate easily interpreted data visualizations in the context of disease; and 5) demonstrate RealSeq-RF technology’s to excel in detection of statistically significant differences in RFs contents between healthy donors and MCD panel plasma samples. We expect that RealSeq-RF technology will be widely adopted for research and potentially disruptive diagnostic applications.