The training of deep neural networks is a high-dimension optimization problem with respect to the loss function of a model. Unfortunately, these functions are of high dimension and non-convex and hence difficult to characterize. In this paper, we empirically investigate the geometry of the loss functions for state-of-the-art networks with multiple stochastic optimization methods. We do this through several experiments that are visualized on polygons to understand how and when these stochastic optimization methods find minima.

Intracellular levels of the amino acid aspartate are responsive to changes in metabolism in mammalian cells and can correspondingly alter cell function, highlighting the need for robust tools to measure aspartate abundance. However, comprehensive understanding of aspartate metabolism has been limited by the throughput, cost, and static nature of the mass spectrometry (MS)-based measurements that are typically employed to measure aspartate levels. To address these issues, we have developed a green fluorescent protein (GFP)-based sensor of aspartate (jAspSnFR3), where the fluorescence intensity corresponds to aspartate concentration. As a purified protein, the sensor has a 20-fold increase in fluorescence upon aspartate saturation, with dose-dependent fluorescence changes covering a physiologically relevant aspartate concentration range and no significant off target binding. Expressed in mammalian cell lines, sensor intensity correlated with aspartate levels measured by MS and could resolve temporal changes in intracellular aspartate from genetic, pharmacological, and nutritional manipulations. These data demonstrate the utility of jAspSnFR3 and highlight the opportunities it provides for temporally resolved and high-throughput applications of variables that affect aspartate levels.

Ends-out gene targeting allows seamless replacement of endogenous genes with engineered DNA fragments by homologous recombination, thus creating designer "genes" in the endogenous locus. Conventional gene targeting in Drosophila involves targeting with the preintegrated donor DNA in the larval primordial germ cells. Here we report G: ene targeting during O: ogenesis with L: ethality I: nhibitor and C: RISPR/Cas (Golic+), which improves on all major steps in such transgene-based gene targeting systems. First, donor DNA is integrated into precharacterized attP sites for efficient flip-out. Second, FLP, I-SceI, and Cas9 are specifically expressed in cystoblasts, which arise continuously from female germline stem cells, thereby providing a continual source of independent targeting events in each offspring. Third, a repressor-based lethality selection is implemented to facilitate screening for correct targeting events. Altogether, Golic+ realizes high-efficiency ends-out gene targeting in ovarian cystoblasts, which can be readily scaled up to achieve high-throughput genome editing.

To establish functional connectivity between two candidate neurons that might form a circuit element, a common approach is to activate an optogenetic tool such as Chrimson in the candidate pre-synaptic neuron and monitor fluorescence of the calcium-sensitive indicator GCaMP in a candidate post-synaptic neuron. While performing such experiments, we found that low levels of leaky Chrimson expression can lead to strong artifactual GCaMP signals in presumptive postsynaptic neurons even when Chrimson is not intentionally expressed in any particular neurons. Withholding all-trans retinal, the chromophore required as a co-factor for Chrimson response to light, eliminates GCaMP signal but does not provide an experimental control for leaky Chrimson expression. Leaky Chrimson expression appears to be an inherent feature of current Chrimson transgenes, since artifactual connectivity was detected with Chrimson transgenes integrated into three different genomic locations (two insertions tested in larvae; a third insertion tested in the adult fly). These false-positive signals may complicate the interpretation of functional connectivity experiments. We illustrate how a no-Gal4 negative control improves interpretability of functional connectivity assays. We also propose a simple but effective procedure to identify experimental conditions that minimize potentially incorrect interpretations caused by leaky Chrimson expression.

Zebrafish larvae are used to model the pathogenesis of multiple bacteria. This transparent model offers the unique advantage of allowing quantification of fluorescent bacterial burdens (fluorescent pixel counts: FPC) in vivo by facile microscopical methods, replacing enumeration of bacteria using time-intensive plating of lysates on bacteriological media. Accurate FPC measurements require laborious manual image processing to mark the outside borders of the animals so as to delineate the bacteria inside the animals from those in the culture medium that they are in. Here, we have developed an automated ImageJ/Fiji-based macro that accurately detect the outside borders of Mycobacterium marinum-infected larvae.

Reference phylogenies are crucial for providing a taxonomic framework for interpretation of marker gene and metagenomic surveys, which continue to reveal novel species at a remarkable rate. Greengenes is a dedicated full-length 16S rRNA gene database that provides users with a curated taxonomy based on de novo tree inference. We developed a ’taxonomy to tree’ approach for transferring group names from an existing taxonomy to a tree topology, and used it to apply the Greengenes, National Center for Biotechnology Information (NCBI) and cyanoDB (Cyanobacteria only) taxonomies to a de novo tree comprising 408 315 sequences. We also incorporated explicit rank information provided by the NCBI taxonomy to group names (by prefixing rank designations) for better user orientation and classification consistency. The resulting merged taxonomy improved the classification of 75% of the sequences by one or more ranks relative to the original NCBI taxonomy with the most pronounced improvements occurring in under-classified environmental sequences. We also assessed candidate phyla (divisions) currently defined by NCBI and present recommendations for consolidation of 34 redundantly named groups. All intermediate results from the pipeline, which includes tree inference, jackknifing and transfer of a donor taxonomy to a recipient tree (tax2tree) are available for download. The improved Greengenes taxonomy should provide important infrastructure for a wide range of megasequencing projects studying ecosystems on scales ranging from our own bodies (the Human Microbiome Project) to the entire planet (the Earth Microbiome Project). The implementation of the software can be obtained from http://sourceforge.net/projects/tax2tree/.

To pursue a more mechanistic understanding of the neural control of behavior, many neuroethologists study animal behavior in controlled laboratory environments. One popular approach is to measure the movements of restrained animals while presenting controlled sensory stimulation. This approach is especially powerful when applied to genetic model organisms, such as , where modern genetic tools enable unprecedented access to the nervous system for activity monitoring or targeted manipulation. While there is a long history of measuring the behavior of body- and head-fixed insects walking on an air-supported ball, the methods typically require complex setups with many custom components. Here we present a compact, simplified setup for these experiments that achieves high-performance at low cost. The simplified setup integrates existing hardware and software solutions with new component designs. We replaced expensive optomechanical and custom machined components with off-the-shelf and 3D-printed parts, and built the system around a low-cost camera that achieves 180 Hz imaging and an inexpensive tablet computer to present view-angle-corrected stimuli updated through a local network. We quantify the performance of the integrated system and characterize the visually guided behavior of flies in response to a range of visual stimuli. In this paper, we thoroughly document the improved system; the accompanying repository incorporates CAD files, parts lists, source code, and detailed instructions. We detail a complete ~$300 system, including a cold-anesthesia tethering stage, that is ideal for hands-on teaching laboratories. This represents a nearly 50-fold cost reduction as compared to a typical system used in research laboratories, yet is fully featured and yields excellent performance. We report the current state of this system, which started with a 1-day teaching lab for which we built seven parallel setups and continues toward a setup in our lab for larger-scale analysis of visual-motor behavior in flies. Because of the simplicity, compactness, and low cost of this system, we believe that high-performance measurements of tethered insect behavior should now be widely accessible and suitable for integration into many systems. This access enables broad opportunities for comparative work across labs, species, and behavioral paradigms.

Across the cell cycle, mitochondrial dynamics are regulated by a cycling wave of actin polymerization/depolymerization. In metaphase, this wave induces actin comet tails on mitochondria that propel these organelles to drive spatial mixing, resulting in their equitable inheritance by daughter cells. In contrast, during interphase the cycling actin wave promotes localized mitochondrial fission. Here, we identify the F-actin nucleator/elongator FMNL1 as a positive regulator of the wave. FMNL1-depleted cells exhibit decreased mitochondrial polarization, decreased mitochondrial oxygen consumption, and increased production of reactive oxygen species. Accompanying these changes is a loss of hetero-fusion of wave-fragmented mitochondria. Thus, we propose that the interphase actin wave maintains mitochondrial homeostasis by promoting mitochondrial content mixing. Finally, we investigate the mechanistic basis for the observation that the wave drives mitochondrial motility in metaphase but mitochondrial fission in interphase. Our data indicate that when the force of actin polymerization is resisted by mitochondrial tethering to microtubules, as in interphase, fission results.

Microinfusions of drugs directly into the central nervous system of awake animals represent a widely used means of unravelling brain functions related to behaviour. However, current approaches generally use tethered liquid infusion systems and a syringe pump to deliver drugs into the brain, which often interfere with behaviour. We address this shortfall with a miniaturised electronically-controlled drug delivery system (20 × 17.5 × 5 mm³) designed to be skull-mounted in rats. The device features a micropump connected to two 8-mm-long silicon microprobes with a cross section of 250 × 250 μm² and integrated fluid microchannels. Using an external electronic control unit, the device allows infusion of 16 metered doses (0.25 μL each, 8 per silicon shaft). Each dosage requires 3.375 Ws of electrical power making the device additionally compatible with state-of-the-art wireless headstages. A dosage precision of 0.25 ± 0.01 μL was determined in vitro before in vivo tests were carried out in awake rats. No passive leakage from the loaded devices into the brain could be detected using methylene blue dye. Finally, the device was used to investigate the effects of the NMDA-receptor antagonist 3-((R)-2-Carboxypiperazin-4-yl)-propyl-1-phosphonic acid, (R)-CPP, administered directly into the prefrontal cortex of rats during performance on a task to assess visual attention and impulsivity. In agreement with previous findings using conventional tethered infusion systems, acute (R)-CPP administration produced a marked increase in impulsivity.

Olfactory stimuli are detected by over 1,000 odorant receptors in mice, with each receptor being mapped to specific glomeruli in the olfactory bulb. The trace amine-associated receptors (TAARs) are a small family of evolutionarily conserved olfactory receptors whose contribution to olfaction remains enigmatic. Here, we show that a majority of the TAARs are mapped to a discrete subset of glomeruli in the dorsal olfactory bulb of the mouse. This TAAR projection is distinct from the previously described class I and class II domains, and is formed by a sensory neuron population that is restricted to express TAAR genes prior to choice. We also show that the dorsal TAAR glomeruli are selectively activated by amines at low concentrations. Our data uncover a hard-wired, parallel input stream in the main olfactory pathway that is specialized for the detection of volatile amines.