Machine learning for modeling animal movement.

Animal movement drives important ecological processes such as migration and the spread of infectious disease. Current approaches to modeling animal tracking data focus on parametric models used to understand environmental effects on movement behavior and to fill in missing tracking data. Machine Learning and Deep learning algorithms are powerful and flexible predictive modeling tools but have rarely been applied to animal movement data. In this study we present a general framework for predicting animal movement that is a combination of two steps: first predicting movement behavioral states and second predicting the animal's velocity. We specify this framework at the individual level as well as for collective movement. We use Random Forests, Neural and Recurrent Neural Networks to compare performance predicting one step ahead as well as long range simulations. We compare results against a custom constructed Stochastic Differential Equation (SDE) model. We apply this approach to high resolution ant movement data. We found that the individual level Machine Learning and Deep Learning methods outperformed the SDE model for one step ahead prediction. The SDE model did comparatively better at simulating long range movement behaviour. Of the Machine Learning and Deep Learning models the Long Short Term Memory (LSTM) individual level model did best at long range simulations. We also applied the Random Forest and LSTM individual level models to model gull migratory movement to demonstrate the generalizability of this framework. Machine Learning and deep learning models are easier to specify compared to traditional parametric movement models which can have restrictive assumptions. However, machine learning and deep learning models are less interpretable than parametric movement models. The type of model used should be determined by the goal of the study, if the goal is prediction, our study provides evidence that machine learning and deep learning models could be useful tools.

Nrf2 mitigates LRRK2- and α-synuclein-induced neurodegeneration by modulating proteostasis.

Mutations in leucine-rich repeat kinase 2 (LRRK2) and α-synuclein lead to Parkinson's disease (PD). Disruption of protein homeostasis is an emerging theme in PD pathogenesis, making mechanisms to reduce the accumulation of misfolded proteins an attractive therapeutic strategy. We determined if activating nuclear factor erythroid 2-related factor (Nrf2), a potential therapeutic target for neurodegeneration, could reduce PD-associated neuron toxicity by modulating the protein homeostasis network. Using a longitudinal imaging platform, we visualized the metabolism and location of mutant LRRK2 and α-synuclein in living neurons at the single-cell level. Nrf2 reduced PD-associated protein toxicity by a cell-autonomous mechanism that was time-dependent. Furthermore, Nrf2 activated distinct mechanisms to handle different misfolded proteins. Nrf2 decreased steady-state levels of α-synuclein in part by increasing α-synuclein degradation. In contrast, Nrf2 sequestered misfolded diffuse LRRK2 into more insoluble and homogeneous inclusion bodies. By identifying the stress response strategies activated by Nrf2, we also highlight endogenous coping responses that might be therapeutically bolstered to treat PD.

A three-groups model for high-throughput survival screens.

Amyotrophic lateral sclerosis (ALS) is a neurodegenerative condition characterized by the progressive deterioration of motor neurons in the cortex and spinal cord. Using an automated robotic microscope platform that enables the longitudinal tracking of thousands of single neurons, we examine the effects a large library of compounds on modulating the survival of primary neurons expressing a mutation known to cause ALS. The goal of our analysis is to identify the few potentially beneficial compounds among the many assayed, the vast majority of which do not extend neuronal survival. This resembles the large-scale simultaneous inference scenario familiar from microarray analysis, but transferred to the survival analysis setting due to the novel experimental setup. We apply a three-component mixture model to censored survival times of thousands of individual neurons subjected to hundreds of different compounds. The shrinkage induced by our model significantly improves performance in simulations relative to performing treatment-wise survival analysis and subsequent multiple testing adjustment. Our analysis identified compounds that provide insight into potential novel therapeutic strategies for ALS.

Proteostasis of polyglutamine varies among neurons and predicts neurodegeneration.

In polyglutamine (polyQ) diseases, only certain neurons die, despite widespread expression of the offending protein. PolyQ expansion may induce neurodegeneration by impairing proteostasis, but protein aggregation and toxicity tend to confound conventional measurements of protein stability. Here, we used optical pulse labeling to measure effects of polyQ expansions on the mean lifetime of a fragment of huntingtin, the protein that causes Huntington's disease, in living neurons. We show that polyQ expansion reduced the mean lifetime of mutant huntingtin within a given neuron and that the mean lifetime varied among neurons, indicating differences in their capacity to clear the polypeptide. We found that neuronal longevity is predicted by the mean lifetime of huntingtin, as cortical neurons cleared mutant huntingtin faster and lived longer than striatal neurons. Thus, cell type-specific differences in turnover capacity may contribute to cellular susceptibility to toxic proteins, and efforts to bolster proteostasis in Huntington's disease, such as protein clearance, could be neuroprotective.

A HIERARCHICAL MAX-STABLE SPATIAL MODEL FOR EXTREME PRECIPITATION.

Extreme environmental phenomena such as major precipitation events manifestly exhibit spatial dependence. Max-stable processes are a class of asymptotically-justified models that are capable of representing spatial dependence among extreme values. While these models satisfy modeling requirements, they are limited in their utility because their corresponding joint likelihoods are unknown for more than a trivial number of spatial locations, preventing, in particular, Bayesian analyses. In this paper, we propose a new random effects model to account for spatial dependence. We show that our specification of the random effect distribution leads to a max-stable process that has the popular Gaussian extreme value process (GEVP) as a limiting case. The proposed model is used to analyze the yearly maximum precipitation from a regional climate model.

Identifying polyglutamine protein species in situ that best predict neurodegeneration.

Polyglutamine (polyQ) stretches exceeding a threshold length confer a toxic function to proteins that contain them and cause at least nine neurological disorders. The basis for this toxicity threshold is unclear. Although polyQ expansions render proteins prone to aggregate into inclusion bodies, this may be a neuronal coping response to more toxic forms of polyQ. The exact structure of these more toxic forms is unknown. Here we show that the monoclonal antibody 3B5H10 recognizes a species of polyQ protein in situ that strongly predicts neuronal death. The epitope selectively appears among some of the many low-molecular-weight conformational states assumed by expanded polyQ and disappears in higher-molecular-weight aggregated forms, such as inclusion bodies. These results suggest that protein monomers and possibly small oligomers containing expanded polyQ stretches can adopt a conformation that is recognized by 3B5H10 and is toxic or closely related to a toxic species.

Quantitative relationships between huntingtin levels, polyglutamine length, inclusion body formation, and neuronal death provide novel insight into huntington's disease molecular pathogenesis.

An expanded polyglutamine (polyQ) stretch in the protein huntingtin (htt) induces self-aggregation into inclusion bodies (IBs) and causes Huntington's disease (HD). Defining precise relationships between early observable variables and neuronal death at the molecular and cellular levels should improve our understanding of HD pathogenesis. Here, we used an automated microscope that tracks thousands of neurons individually over their entire lifetime to quantify interconnected relationships between early variables, such as htt levels, polyQ length, and IB formation, and neuronal death in a primary striatal model of HD. The resulting model revealed that mutant htt increases the risk of death by tonically interfering with homeostatic coping mechanisms rather than producing accumulated damage to the neuron, htt toxicity is saturable, the rate-limiting steps for inclusion body formation and death can be traced to different conformational changes in monomeric htt, and IB formation reduces the impact of the starting levels of htt of a neuron on its risk of death. Finally, the model that emerges from our quantitative measurements places critical limits on the potential mechanisms by which mutant htt might induce neurodegeneration, which should help direct future research.