Rapid Acute Leukemia Genomic Profiling with CRISPR enrichment and Real-time long-read sequencing

点击翻译按钮获取中文摘要

PROJECT SUMMARY The parent award (1R21CA280520-01) aims to develop fast and cost-effective methods of detecting Acute Myeloid Leukemia (AML) fusions and mutations to improve clinical outcomes for patients. The parent award builds on a CRISPR-based single molecule long-read sequencing assay (CSRL) to detect clinically relevant mutated genes in AML by including additional prognostic and predictive markers. The improved assay will be validated using known cell line mixes and patient samples. Aim 1 of the parent R21 will also create bioinformatics workflows to enable same day data analysis using the open-source Biodepot-workflow-builder (Bwb) platform. The Bwb platform will be extended to detect fusion and DNA variants. Aim 2 of the parent R21 will compare our assay performance to established clinical grade tests on archival specimens, and evaluate clinical impact of phasing data and fusion breakpoints on prognosis and treatment response. Aim 3 of the parent R21 will demonstrate the clinical impact of our molecular profiling assay and bioinformatics platform through a pilot study on samples from the clinic, ease of same day testing and reporting, and study of the impacts that same day results can potentially have on patient treatment. Our CSRL assay, combined with new bioinformatics approaches, will establish potential use in the clinic for an ultrarapid same-day informative molecular profiling assay for AML to guide oncologists for immediate therapy implementations. In this supplement, we will employ software engineering best practices to extend the Bwb platform to enable execution of modules on different platforms that are specified at runtime. Bioinformatics workflows will be able to simultaneously utilize different resources and leverage the enhanced security and reduced data transfer of local devices while benefiting from the scalability of the cloud. This hybrid cloud computing approach would be beneficial for the parent R21 as steps involving large files (basecalling and alignment) could be done locally on a GPU enabled laptop, server, or dedicated device resulting in the transfer of smaller files to the cloud for computationally intensive deep-learning based variant calling. Bwb currently has specialized workflows with modules that run simultaneously on the cloud and locally. The supplement will allow users to specify the execution platform for modules at runtime without the need to re-write the workflow. We will support a full-time research software engineer and add a new collaborator with expertise in software engineering and cloud computing. This hybrid cloud implementation will be applied and benchmarked using data generated in the parent award and the test suite will be saved and distributed as part of the workflow. We will also provide a single sign-on service to simplify the additional authentication needed for execution on multiple platforms. All software developed will be public and open-source. These improvements will provide a robust open-source platform for other multi-omics and imaging data workflows to leverage emerging hybrid cloud and edge computing technologies.