Provable Convex Co-clustering of Tensors.

Cluster analysis is a fundamental tool for pattern discovery of complex heterogeneous data. Prevalent clustering methods mainly focus on vector or matrix-variate data and are not applicable to general-order tensors, which arise frequently in modern scientific and business applications. Moreover, there is a gap between statistical guarantees and computational efficiency for existing tensor clustering solutions due to the nature of their non-convex formulations. In this work, we bridge this gap by developing a provable convex formulation of tensor co-clustering. Our convex co-clustering (CoCo) estimator enjoys stability guarantees and its computational and storage costs are polynomial in the size of the data. We further establish a non-asymptotic error bound for the CoCo estimator, which reveals a surprising "blessing of dimensionality" phenomenon that does not exist in vector or matrix-variate cluster analysis. Our theoretical findings are supported by extensive simulated studies. Finally, we apply the CoCo estimator to the cluster analysis of advertisement click tensor data from a major online company. Our clustering results provide meaningful business insights to improve advertising effectiveness.

EFFECT OF A NUTRIENT ENEMA ON SERUM NUTRIENT CONCENTRATIONS IN WHITE-SPOTTED BAMBOO SHARKS (CHILOSCYLLIUM PLAGIOSUM).

Ill and anorectic captive sharks present a unique challenge for husbandry and veterinary staff. Providing adequate fluid and nutritional support to sharks while minimizing handling remains difficult. This study aimed to evaluate the ability of a nutrient enema to alter blood analyte concentrations. Thirty-six healthy, fasted white-spotted bamboo sharks (Chiloscyllium plagiosum) were enrolled in the study with 18 sharks receiving a nutrient enema and 18 sharks receiving a non-nutrient saline enema. The metabolic state of sharks was evaluated via measurement of blood glucose, blood urea nitrogen, and ß-hydroxybutyrate as well as other serum biochemistry parameters. Changes in sodium, chloride, calcium, ß-hydroxybutyrate, glucose, total protein, and triglyceride concentrations were seen across time in both groups. Blood glucose absolute concentrations and changes over time differed between the nutrient and nonnutrient groups. This pilot study indicates that it is possible to influence the glucose metabolism of healthy sharks via nutrient enema. Further study is needed to better understand potential therapeutics for ill and anorectic sharks.

Improving Undergraduate Life Science Education for the Biosciences Workforce: Overcoming the Disconnect between Educators and Industry.

The BioHealth Capital Region (Maryland, Virginia, and Washington, DC; BHCR) is flush with colleges and universities training students in science, technology, engineering, and mathematics disciplines and has one of the most highly educated workforces in the United States. However, current educational approaches and business recruitment tactics are not drawing sufficient talent to sustain the bioscience workforce pipeline. Surveys conducted by the Mid-Atlantic Biology Research and Career Network identified a disconnect between stakeholders who are key to educating, training, and hiring college and university graduates, resulting in several impediments to workforce development in the BHCR: 1) students are underinformed or unaware of bioscience opportunities before entering college and remain so at graduation; 2) students are not job ready at the time of graduation; 3) students are mentored to pursue education beyond what is needed and are therefore overqualified (by degree) for most of the available jobs in the region; 4) undergraduate programs generally lack any focus on workforce development; and 5) few industry-academic partnerships with undergraduate institutions exist in the region. The reality is that these issues are neither surprising nor restricted to the BHCR. Recommendations are presented to facilitate improvement in the preparation of graduates for today's bioscience industries throughout the United States.

Recovery of horses from general anesthesia after induction with propofol and ketamine versus midazolam and ketamine.

OBJECTIVE To evaluate quality of recovery from general anesthesia in horses after induction with propofol and ketamine versus midazolam and ketamine. DESIGN Prospective randomized crossover study. ANIMALS 6 healthy adult horses. PROCEDURES Horses were premedicated with xylazine (1.0 mg/kg [0.45 mg/lb], IV), and general anesthesia was induced with midazolam (0.1 mg/kg [0.045 mg/lb], IV) or propofol (0.5 mg/kg [0.23 mg/lb], IV), followed by ketamine (3.0 mg/kg [1.36 mg/lb], IV). Horses were endotracheally intubated, and anesthesia was maintained with isoflurane. After 60 minutes, horses were given romifidine (0.02 mg/kg [0.009 mg/lb], IV) and allowed to recover unassisted. Times to first movement, sternal recumbency, and standing and the number of attempts to stand were recorded. Plasma concentrations of propofol or midazolam were measured following induction and immediately before recovery. Recovery quality was scored by 3 graders with a recovery rubric and a visual analog scale. RESULTS Number of attempts to stand was significantly lower when horses received propofol (median, 2; range, 1 to 3) than when they received midazolam (median, 7.5; range, 3 to 16). For both the recovery rubric and visual analog scale, recovery quality was significantly better when horses received propofol than when they received midazolam. Plasma drug concentration at recovery, as a percentage of the concentration at induction, was significantly lower when horses received propofol than when they received midazolam. CONCLUSIONS AND CLINICAL RELEVANCE Results suggested that for horses undergoing short (ie, 60 minutes) periods of general anesthesia, recovery quality may be better following induction with propofol and ketamine, compared with midazolam and ketamine.

Algorithms for Fitting the Constrained Lasso.

We compare alternative computing strategies for solving the constrained lasso problem. As its name suggests, the constrained lasso extends the widely-used lasso to handle linear constraints, which allow the user to incorporate prior information into the model. In addition to quadratic programming, we employ the alternating direction method of multipliers (ADMM) and also derive an efficient solution path algorithm. Through both simulations and benchmark data examples, we compare the different algorithms and provide practical recommendations in terms of efficiency and accuracy for various sizes of data. We also show that, for an arbitrary penalty matrix, the generalized lasso can be transformed to a constrained lasso, while the converse is not true. Thus, our methods can also be used for estimating a generalized lasso, which has wide-ranging applications. Code for implementing the algorithms is freely available in both the Matlab toolbox SparseReg and the Julia package ConstrainedLasso. Supplementary materials for this article are available online.

PROPOFOL AS AN IMMERSION ANESTHETIC AND IN A MINIMUM ANESTHETIC CONCENTRATION (MAC) REDUCTION MODEL IN GOLDFISH (CARASSIUS AURATUS).

Propofol is a novel immersion anesthetic in goldfish ( Carassius auratus ). Objectives were to characterize propofol as an anesthetic and assess its suitability in a minimum anesthetic concentration (MAC) reduction model. Using a crossover design, eight goldfish were submerged in 1, 5, or 10 mg/L propofol. Data included induction time, recovery time, heart rate, opercular rate, and response to supramaximal stimulation. Baseline MAC (Dixon's up-and-down method) was determined, and 15 fish were anesthetized with propofol on 4 consecutive days with MAC determination on the fifth day, weekly, for 1 mo. Using a crossover design, MAC of propofol (n = 15) was determined 1 hr following administration of i.m. butorphanol 0.05, 0.5, and 1 mg/kg, dexmedetomidine 0.01, 0.02, and 0.04 mg/kg, ketoprofen 0.5, 1, and 2 mg/kg, morphine 5, 10, and 15 mg/kg, or saline 1 ml/kg. Comparisons were performed with Wilcoxon signed-rank tests (P < 0.05) and Tango's score confidence interval. Propofol at 1 mg/L did not produce anesthesia. Induction time with 10 mg/L (112, 84-166 s) was faster than 5 mg/L (233, 150-289 s; P = 0.0078). Heart and opercular rates for 5 and 10 mg/L were 36 (24-72) beats/min, 58 (44-68) operculations/min and 39 (20-48) beats/min, 57 (48-80) operculations/min, respectively. Recovery time was 249 (143-396) s and 299 (117-886) s with 5 and 10 mg/L, respectively. Response to supramaximal stimulation was not significantly different with 5 mg/L (1/8) compared with 10 mg/L (0/8). Baseline and weekly MAC following daily exposure was 8.4 and 9.0, 8.1, 8.1, and 8.7 mg/L, respectively. MAC reduction was no more than 8% following any drug or dosage. Propofol at 5 and 10 mg/L produced anesthesia, and anesthetic needs were similar following repeated exposure. Propofol was not suitable to test MAC reduction in goldfish in this study.