Abstract WP288: A Portable and Rapid Stroke Imaging and Classification Device

Introduction: In stroke management, time to treatment is directly proportional to morbidity and mortality. The time lapse between arriving at a hospital and receiving treatment can easily exceed one hour, with most of this time being spent waiting for and operating the CT machine. Imaging is necessary prior to treatment to differentiate between ischemic and hemorrhagic stroke, which boast vastly different treatments. Here, we describe a novel portable sensor that can classify stroke and produce images in minutes. Objective: To create of a compact, noninvasive device that can classify stroke and produce an image to localize hemorrhages in minutes. Methods: Multiple inductive-damping sensors with varying magnetic field strengths were built in house, with the largest sensor 11cm in diameter. Sixteen gelatin brain models with identical electrical properties to live brain tissue were developed and placed within plastic skull replicas. Saline was diluted to the conductivity of blood and placed within the brain by a third party to simulate a hemorrhage. Bleed sizes ranged from 20-200mL. Individuals scanning were blinded to bleed location, and sensors were tangentially rotated around the skull models to localize blood. Data was also used to create MRI-style images using MATLAB. Results: Our novel sensor accurately predicted the location of bleeds in all 16 experiments. The lower limit of volume detection was found to be 25mL. The maximal blood detection range of the sensors was found to be 5cm from the skull, and hemorrhage detection and image processing took 2.43 minutes on average (attached figure, bottom images are actual bleed location, top images are predicted bleed location from sensor). Conclusion: We demonstrate the feasibility of a device that may reduce the high morbidity and mortality associated with stroke by providing rapid bedside diagnostics and reduced time to treatment. Further animal and human experiments are necessary to fully establish the efficacy of this device.

Toward fast and accurate SNP genotyping from whole genome sequencing data for bedside diagnostics

MotivationGenotyping a set of variants from a database is an important step for identifying known genetic traits and disease related variants within an individual. The growing size of variant databases as well as the high depth of sequencing data pose an efficiency challenge. In clinical applications, where time is crucial, alignment-based methods are often not fast enough. To fill the gap, Shajii et al. (2016) propose LAVA, an alignment-free genotyping method which is able to more quickly genotype SNPs; however, there remains large room for improvements in running time and accuracy.ResultsWe present the VarGeno method for SNP genotyping from lllumina whole genome sequencing data. VarGeno builds upon LAVA by improving the speed of k-mer querying as well as the accuracy of the genotyping strategy. We evaluate VarGeno on several read datasets using different genotyping SNP lists. VarGeno performs 7-13 times faster than LAVA with similar memory usage, while improving accuracy.AvailabilityVarGeno is freely available at: https://github.com/medvedevgroup/vargeno.