Search for: All records

For stationary time series with regularly varying marginal distributions, an important problem is to estimate the associated tail index which characterizes the power‐law behavior of the tail distribution. For this, various results have been developed for independent data and certain types of dependent data. In this article, we consider the problem of tail index estimation under a recently proposed notion of serial tail dependence called the tail adversarial stability. Using the technique of adversarial innovation coupling and a martingale approximation scheme, we establish the consistency and central limit theorem of the tail index estimator for a general class of tail dependent time series. Based on the asymptotic normal distribution from the obtained central limit theorem, we further consider an application to cluster a large number of regularly varying time series based on their tail indices by using a robust mixture algorithm. The results are illustrated using numerical examples including Monte Carlo simulations and a real data analysis. 

Virus-like particles (VLPs) have been proposed as an attractive tool in SARS-CoV-2 vaccine development, both as (1) a vaccine candidate with high immunogenicity and low reactogenicity and (2) a substitute for live virus in functional and neutralization assays. Though multiple SARS-CoV-2 VLP designs have already been explored in Sf9 insect cells, a key parameter ensuring VLPs are a viable platform is the VLP spike yield (i.e., spike protein content in VLP), which has largely been unreported. In this study, we show that the common strategy of producing SARS-CoV-2 VLPs by expressing spike protein in combination with the native coronavirus membrane and/or envelope protein forms VLPs, but at a critically low spike yield (~0.04–0.08 mg/L). In contrast, fusing the spike ectodomain to the influenza HA transmembrane domain and cytoplasmic tail and co-expressing M1 increased VLP spike yield to ~0.4 mg/L. More importantly, this increased yield translated to a greater VLP spike antigen density (~96 spike monomers/VLP) that more closely resembles that of native SARS-CoV-2 virus (~72–144 Spike monomers/virion). Pseudotyping further allowed for production of functional alpha (B.1.1.7), beta (B.1.351), delta (B.1.617.2), and omicron (B.1.1.529) SARS-CoV-2 VLPs that bound to the target ACE2 receptor. Finally, we demonstrated the utility of pseudotyped VLPs to test neutralizing antibody activity using a simple, acellular ELISA-based assay performed at biosafety level 1 (BSL-1). Taken together, this study highlights the advantage of pseudotyping over native SARS-CoV-2 VLP designs in achieving higher VLP spike yield and demonstrates the usefulness of pseudotyped VLPs as a surrogate for live virus in vaccine and therapeutic development against SARS-CoV-2 variants. 

Key-value (KV) software has proven useful to a wide variety of applications including analytics, time-series databases, and distributed file systems. To satisfy the requirements of diverse workloads, KV stores have been carefully tailored to best match the performance characteristics of underlying solid-state block devices. Emerging KV storage device is a promising technology for both simplifying the KV software stack and improving the performance of persistent storage-based applications. However, while providing fast, predictable put and get operations, existing KV storage devices don’t natively support range queries which are critical to all three types of applications described above. In this paper, we present KVRangeDB, a software layer that enables processing range queries for existing hash-based KV solid-state disks (KVSSDs). As an effort to adapt to the performance characteristics of emerging KVSSDs, KVRangeDB implements log-structured merge tree key index that reduces compaction I/O, merges keys when possible, and provides separate caches for indexes and values. We evaluated the KVRangeDB under a set of representative workloads, and compared its performance with two existing database solutions: a Rocksdb variant ported to work with the KVSSD, and Wisckey, a key-value database that is carefully tuned for conventional block devices. On filesystem aging workloads, KVRangeDB outperforms Wisckey by 23.7x in terms of throughput and reduce CPU usage and external write amplifications by 14.3x and 9.8x, respectively.