Low dimensional measurement of vowels using machine perception.

A method is presented for combining the feature extraction power of neural networks with model based dimensionality reduction to produce linguistically motivated low dimensional measurements of sounds. This method works by first training a convolutional neural network (CNN) to predict linguistically relevant category labels from the spectrograms of sounds. Then, idealized models of these categories are defined as probability distributions in a low dimensional measurement space with locations chosen to reproduce, as far as possible, the perceptual characteristics of the CNN. To measure a sound, the point is found in the measurement space for which the posterior probability distribution over categories in the idealized model most closely matches the category probabilities output by the CNN for that sound. In this way, the feature learning power of the CNN is used to produce low dimensional measurements. This method is demonstrated using monophthongal vowel categories to train this CNN and produce measurements in two dimensions. It is also shown that the perceptual characteristics of this CNN are similar to those of human listeners.

Optimizing Crop Water Use for Drought and Climate Change Adaptation Requires a Multi-Scale Approach.

Selection criteria that co-optimize water use efficiency and yield are needed to promote plant productivity in increasingly challenging and variable drought scenarios, particularly dryland cereals in the semi-arid tropics. Optimizing water use efficiency and yield fundamentally involves transpiration dynamics, where restriction of maximum transpiration rate helps to avoid early crop failure, while maximizing grain filling. Transpiration restriction can be regulated by multiple mechanisms and involves cross-organ coordination. This coordination involves complex feedbacks and feedforwards over time scales ranging from minutes to weeks, and from spatial scales ranging from cell membrane to crop canopy. Aquaporins have direct effect but various compensation and coordination pathways involve phenology, relative root and shoot growth, shoot architecture, root length distribution profile, as well as other architectural and anatomical aspects of plant form and function. We propose gravimetric phenotyping as an integrative, cross-scale solution to understand the dynamic, interwoven, and context-dependent coordination of transpiration regulation. The most fruitful breeding strategy is likely to be that which maintains focus on the phene of interest, namely, daily and season level transpiration dynamics. This direct selection approach is more precise than yield-based selection but sufficiently integrative to capture attenuating and complementary factors.

Improving Association Studies and Genomic Predictions for Climbing Beans With Data From Bush Bean Populations.

Common bean (Phaseolus vulgaris L.) has two major origins of domestication, Andean and Mesoamerican, which contribute to the high diversity of growth type, pod and seed characteristics. The climbing growth habit is associated with increased days to flowering (DF), seed iron concentration (SdFe), nitrogen fixation, and yield. However, breeding efforts in climbing beans have been limited and independent from bush type beans. To advance climbing bean breeding, we carried out genome-wide association studies and genomic predictions using 1,869 common bean lines belonging to five breeding panels representing both gene pools and all growth types. The phenotypic data were collected from 17 field trials and were complemented with 16 previously published trials. Overall, 38 significant marker-trait associations were identified for growth habit, 14 for DF, 13 for 100 seed weight, three for SdFe, and one for yield. Except for DF, the results suggest a common genetic basis for traits across all panels and growth types. Seven QTL associated with growth habits were confirmed from earlier studies and four plausible candidate genes for SdFe and 100 seed weight were newly identified. Furthermore, the genomic prediction accuracy for SdFe and yield in climbing beans improved up to 8.8% when bush-type bean lines were included in the training population. In conclusion, a large population from different gene pools and growth types across multiple breeding panels increased the power of genomic analyses and provides a solid and diverse germplasm base for genetic improvement of common bean.

Gradual domestication of root traits in the earliest maize from Tehuacán.

Efforts to understand the phenotypic transition that gave rise to maize from teosinte have mainly focused on the analysis of aerial organs, with little insights into possible domestication traits affecting the root system. Archeological excavations in San Marcos cave (Tehuacán, Mexico) yielded two well-preserved 5,300 to 4,970 calibrated y B.P. specimens (SM3 and SM11) corresponding to root stalks composed of at least five nodes with multiple nodal roots and, in case, a complete embryonic root system. To characterize in detail their architecture and anatomy, we used laser ablation tomography to reconstruct a three-dimensional segment of their nodal roots and a scutellar node, revealing exquisite preservation of the inner tissue and cell organization and providing reliable morphometric parameters for cellular characteristics of the stele and cortex. Whereas SM3 showed multiple cortical sclerenchyma typical of extant maize, the scutellar node of the SM11 embryonic root system completely lacked seminal roots, an attribute found in extant teosinte and in two specific maize mutants: root with undetectable meristem1 (rum1) and rootless concerning crown and seminal roots (rtcs). Ancient DNA sequences of SM10­a third San Marcos specimen of equivalent age to SM3 and SM11­revealed the presence of mutations in the transcribed sequence of both genes, offering the possibility for some of these mutations to be involved in the lack of seminal roots of the ancient specimens. Our results indicate that the root system of the earliest maize from Tehuacán resembled teosinte in traits important for maize drought adaptation.

Root traits for low input agroecosystems in Africa: Lessons from three case studies.

In many regions across Africa, agriculture is largely based on low-input and small-holder farming systems that use little inorganic fertilisers and have limited access to irrigation and mechanisation. Improving agricultural practices and developing new cultivars adapted to these environments, where production already suffers from climate change, is a major priority for food security. Here, we illustrate how breeding for specific root traits could improve crop resilience in Africa using three case studies covering very contrasting low-input agroecosystems. We first review how greater basal root whorl number and longer and denser root hairs increased P acquisition efficiency and yield in common bean in South East Africa. We then discuss how water-saving strategies, root hair density and deep root growth could be targeted to improve sorghum and pearl millet yield in West Africa. Finally, we evaluate how breeding for denser root systems in the topsoil and interactions with arbuscular mycorrhizal fungi could be mobilised to optimise water-saving alternate wetting and drying practices in West African rice agroecosystems. We conclude with a discussion on how to evaluate the utility of root traits and how to make root trait selection feasible for breeders so that improved varieties can be made available to farmers through participatory approaches.

Public efforts to reduce disease transmission implied from a spatial game.

One approach to understand people's efforts to reduce disease transmission, is to consider the effect of behaviour on case rates. In this paper we present a spatial infection-reducing game model of public behaviour, formally equivalent to a Hopfield neural network coupled to SIRS disease dynamics. Behavioural game parameters can be precisely calibrated to geographical time series of Covid-19 active case numbers, giving an implied spatial history of behaviour. This is used to investigate the effects of government intervention, quantify behaviour area by area, and measure the effect of wealth on behaviour. We also demonstrate how a delay in people's perception of risk levels can induce behavioural instability, and oscillations in infection rates.

Root metaxylem and architecture phenotypes integrate to regulate water use under drought stress.

At the genus and species level, variation in root anatomy and architecture may interact to affect strategies of drought avoidance. To investigate this idea, root anatomy and architecture of the drought-sensitive common bean (Phaseolus vulgaris) and drought-adapted tepary bean (Phaseolus acutifolius) were analyzed in relation to water use under terminal drought. Intraspecific variation for metaxylem anatomy and axial conductance was found in the roots of both species. Genotypes with high-conductance root metaxylem phenotypes acquired and transpired more water per unit leaf area, shoot mass, and root mass than genotypes with low-conductance metaxylem phenotypes. Interspecific variation in root architecture and root depth was observed where P. acutifolius has a deeper distribution of root length than P. vulgaris. In the deeper-rooted P. acutifolius, genotypes with high root conductance were better able to exploit deep soil water than genotypes with low root axial conductance. Contrastingly, in the shallower-rooted P. vulgaris, genotypes with low root axial conductance had improved water status through conservation of soil moisture for sustained water capture later in the season. These results indicate that metaxylem morphology interacts with root system depth to determine a strategy of drought avoidance and illustrate synergism among architectural and anatomical phenotypes for root function.

An Analysis of Soil Coring Strategies to Estimate Root Depth in Maize (Zea mays) and Common Bean (Phaseolus vulgaris).

A soil coring protocol was developed to cooptimize the estimation of root length distribution (RLD) by depth and detection of functionally important variation in root system architecture (RSA) of maize and bean. The functional-structural model OpenSimRoot was used to perform in silico soil coring at six locations on three different maize and bean RSA phenotypes. Results were compared to two seasons of field soil coring and one trench. Two one-sided T-test (TOST) analysis of in silico data suggests a between-row location 5 cm from plant base (location 3), best estimates whole-plot RLD/D of deep, intermediate, and shallow RSA phenotypes, for both maize and bean. Quadratic discriminant analysis indicates location 3 has ~70% categorization accuracy for bean, while an in-row location next to the plant base (location 6) has ~85% categorization accuracy in maize. Analysis of field data suggests the more representative sampling locations vary by year and species. In silico and field studies suggest location 3 is most robust, although variation is significant among seasons, among replications within a field season, and among field soil coring, trench, and simulations. We propose that the characterization of the RLD profile as a dynamic rhizo canopy effectively describes how the RLD profile arises from interactions among an individual plant, its neighbors, and the pedosphere.

ASTEROID: A New Clinical Stereotest on an Autostereo 3D Tablet.

PURPOSE: To describe a new stereotest in the form of a game on an autostereoscopic tablet computer designed to be suitable for use in the eye clinic and present data on its reliability and the distribution of stereo thresholds in adults. METHODS: Test stimuli were four dynamic random-dot stereograms, one of which contained a disparate target. Feedback was given after each trial presentation. A Bayesian adaptive staircase adjusted target disparity. Threshold was estimated from the mean of the posterior distribution after 20 responses. Viewing distance was monitored via a forehead sticker viewed by the tablet's front camera, and screen parallax was adjusted dynamically so as to achieve the desired retinal disparity. RESULTS: The tablet must be viewed at a distance of greater than ∼35 cm to produce a good depth percept. Log thresholds were roughly normally distributed with a mean of 1.75 log10 arcsec = 56 arcsec and SD of 0.34 log10 arcsec = a factor of 2.2. The standard deviation agrees with previous studies, but ASTEROID thresholds are approximately 1.5 times higher than a similar stereotest on stereoscopic 3D TV or on Randot Preschool stereotests. Pearson correlation between successive tests in same observer was 0.80. Bland-Altman 95% limits of reliability were ±0.64 log10 arcsec = a factor of 4.3, corresponding to an SD of 0.32 log10 arcsec on individual threshold estimates. This is similar to other stereotests and close to the statistical limit for 20 responses. CONCLUSIONS: ASTEROID is reliable, easy, and portable and thus well-suited for clinical stereoacuity measurements. TRANSLATIONAL RELEVANCE: New 3D digital technology means that research-quality psychophysical measurement of stereoacuity is now feasible in the clinic.

Unifying models of dialect spread and extinction using surface tension dynamics.

We provide a unified mathematical explanation of two classical forms of spatial linguistic spread. The wave model describes the radiation of linguistic change outwards from a central focus. Changes can also jump between population centres in a process known as hierarchical diffusion. It has recently been proposed that the spatial evolution of dialects can be understood using surface tension at linguistic boundaries. Here we show that the inclusion of long-range interactions in the surface tension model generates both wave-like spread, and hierarchical diffusion, and that it is surface tension that is the dominant effect in deciding the stable distribution of dialect patterns. We generalize the model to allow population mixing which can induce shrinkage of linguistic domains, or destroy dialect regions from within.

Genome-wide association mapping and agronomic impact of cowpea root architecture.

KEY MESSAGE: Genetic analysis of data produced by novel root phenotyping tools was used to establish relationships between cowpea root traits and performance indicators as well between root traits and Striga tolerance. Selection and breeding for better root phenotypes can improve acquisition of soil resources and hence crop production in marginal environments. We hypothesized that biologically relevant variation is measurable in cowpea root architecture. This study implemented manual phenotyping (shovelomics) and automated image phenotyping (DIRT) on a 189-entry diversity panel of cowpea to reveal biologically important variation and genome regions affecting root architecture phenes. Significant variation in root phenes was found and relatively high heritabilities were detected for root traits assessed manually (0.4 for nodulation and 0.8 for number of larger laterals) as well as repeatability traits phenotyped via DIRT (0.5 for a measure of root width and 0.3 for a measure of root tips). Genome-wide association study identified 11 significant quantitative trait loci (QTL) from manually scored root architecture traits and 21 QTL from root architecture traits phenotyped by DIRT image analysis. Subsequent comparisons of results from this root study with other field studies revealed QTL co-localizations between root traits and performance indicators including seed weight per plant, pod number, and Striga (Striga gesnerioides) tolerance. The data suggest selection for root phenotypes could be employed by breeding programs to improve production in multiple constraint environments.

Birdsong dialect patterns explained using magnetic domains.

The songs and calls of many bird species, like human speech, form distinct regional dialects. We suggest that the process of dialect formation is analogous to the physical process of magnetic domain formation. We take the coastal breeding grounds of the Puget Sound white crowned sparrow as an example. Previous field studies suggest that birds of this species learn multiple songs early in life, and when establishing a territory for the first time, retain one of these dialects in order to match the majority of their neighbors. We introduce a simple lattice model of the process, showing that this matching behavior can produce single dialect domains provided the death rate of adult birds is sufficiently low. We relate death rate to thermodynamic temperature in magnetic materials, and calculate the critical death rate by analogy with the Ising model. Using parameters consistent with the known behavior of these birds we show that coastal dialect domain shapes may be explained by viewing them as low-temperature "stripe states."

Infrequent social interaction can accelerate the spread of a persuasive idea.

We study the spread of a persuasive new idea through a population of continuous-time random walkers in one dimension. The idea spreads via social gatherings involving groups of nearby walkers who act according to a biased "majority rule": After each gathering, the group takes on the new idea if more than a critical fraction 1-ɛ/2<1/2 of them already hold it; otherwise they all reject it. The boundary of a domain where the new idea has taken hold expands as a traveling wave in the density of new idea holders. Our walkers move by Lévy motion, and we compute the wave velocity analytically as a function of the frequency of social gatherings and the exponent of the jump distribution. When this distribution is sufficiently heavy tailed, then, counter to intuition, the idea can propagate faster if social gatherings are held less frequently. When jumps are truncated, a critical gathering frequency can emerge which maximizes propagation velocity. We explore our model by simulation, confirming our analytical results.

Limit cycles and the benefits of a short memory in rock-paper-scissors games.

When playing games in groups, it is an advantage for individuals to have accurate statistical information on the strategies of their opponents. Such information may be obtained by remembering previous interactions. We consider a rock-paper-scissors game in which agents are able to recall their last m interactions, used to estimate the behavior of their opponents. At critical memory length, a Hopf bifurcation leads to the formation of stable limit cycles. In a mixed population, agents with longer memories have an advantage, provided the system has a stable fixed point, and there is some asymmetry in the payoffs of the pure strategies. However, at a critical concentration of long memory agents, the appearance of limit cycles destroys their advantage. By introducing population dynamics that favors successful agents, we show that the system evolves toward the bifurcation point.

Digital imaging of root traits (DIRT): a high-throughput computing and collaboration platform for field-based root phenomics.

BACKGROUND: Plant root systems are key drivers of plant function and yield. They are also under-explored targets to meet global food and energy demands. Many new technologies have been developed to characterize crop root system architecture (CRSA). These technologies have the potential to accelerate the progress in understanding the genetic control and environmental response of CRSA. Putting this potential into practice requires new methods and algorithms to analyze CRSA in digital images. Most prior approaches have solely focused on the estimation of root traits from images, yet no integrated platform exists that allows easy and intuitive access to trait extraction and analysis methods from images combined with storage solutions linked to metadata. Automated high-throughput phenotyping methods are increasingly used in laboratory-based efforts to link plant genotype with phenotype, whereas similar field-based studies remain predominantly manual low-throughput. DESCRIPTION: Here, we present an open-source phenomics platform "DIRT", as a means to integrate scalable supercomputing architectures into field experiments and analysis pipelines. DIRT is an online platform that enables researchers to store images of plant roots, measure dicot and monocot root traits under field conditions, and share data and results within collaborative teams and the broader community. The DIRT platform seamlessly connects end-users with large-scale compute "commons" enabling the estimation and analysis of root phenotypes from field experiments of unprecedented size. CONCLUSION: DIRT is an automated high-throughput computing and collaboration platform for field based crop root phenomics. The platform is accessible at http://www.dirt.iplantcollaborative.org/ and hosted on the iPlant cyber-infrastructure using high-throughput grid computing resources of the Texas Advanced Computing Center (TACC). DIRT is a high volume central depository and high-throughput RSA trait computation platform for plant scientists working on crop roots. It enables scientists to store, manage and share crop root images with metadata and compute RSA traits from thousands of images in parallel. It makes high-throughput RSA trait computation available to the community with just a few button clicks. As such it enables plant scientists to spend more time on science rather than on technology. All stored and computed data is easily accessible to the public and broader scientific community. We hope that easy data accessibility will attract new tool developers and spur creative data usage that may even be applied to other fields of science.

Forgetfulness can help you win games.

We present a simple game model where agents with different memory lengths compete for finite resources. We show by simulation and analytically that an instability exists at a critical memory length, and as a result, different memory lengths can compete and coexist in a dynamical equilibrium. Our analytical formulation makes a connection to statistical urn models, and we show that temperature is mirrored by the agent's memory. Our simple model of memory may be incorporated into other game models with implications that we briefly discuss.

Formation and arrangement of pits by a corrosive gas.

When corroding or otherwise aggressive particles are incident on a surface, pits can form. For example, under certain circumstances rock surfaces that are exposed to salts can form regular tessellating patterns of pits known as "tafoni." We introduce a simple lattice model in which a gas of corrosive particles, described by a discrete, biased diffusion equation, drifts onto a surface. Each gas particle has a fixed probability of being absorbed and causing damage at each contact. The surface is represented by a lattice of strength numbers which reduce after each absorbtion event, with sites being removed when their strength becomes negative. Regular formations of pits arise spontaneously, with each pit having a characteristic trapezoidal geometry determined by the particle bias, absorbtion probability, and surface strength. The formation of this geometry may be understood in terms of a first order partial differential equation and is a consequence of particle concentration gradients which arise in the pits. By viewing pits as particle funnels, we are able to relate the gradient of pit walls to absorbtion probability and particle bias.

Image-based high-throughput field phenotyping of crop roots.

Current plant phenotyping technologies to characterize agriculturally relevant traits have been primarily developed for use in laboratory and/or greenhouse conditions. In the case of root architectural traits, this limits phenotyping efforts, largely, to young plants grown in specialized containers and growth media. Hence, novel approaches are required to characterize mature root systems of older plants grown under actual soil conditions in the field. Imaging methods able to address the challenges associated with characterizing mature root systems are rare due, in part, to the greater complexity of mature root systems, including the larger size, overlap, and diversity of root components. Our imaging solution combines a field-imaging protocol and algorithmic approach to analyze mature root systems grown in the field. Via two case studies, we demonstrate how image analysis can be utilized to estimate localized root traits that reliably capture heritable architectural diversity as well as environmentally induced architectural variation of both monocot and dicot plants. In the first study, we show that our algorithms and traits (including 13 novel traits inaccessible to manual estimation) can differentiate nine maize (Zea mays) genotypes 8 weeks after planting. The second study focuses on a diversity panel of 188 cowpea (Vigna unguiculata) genotypes to identify which traits are sufficient to differentiate genotypes even when comparing plants whose harvesting date differs up to 14 d. Overall, we find that automatically derived traits can increase both the speed and reproducibility of the trait estimation pipeline under field conditions.

Shocks generate crossover behavior in lattice avalanches.

A spatial avalanche model is introduced, in which avalanches increase stability in the regions where they occur. Instability is driven globally by a driving process that contains shocks. The system is typically subcritical, but the shocks occasionally lift it into a near- or supercritical state from which it rapidly retreats due to large avalanches. These shocks leave behind a signature-a distinct power-law crossover in the avalanche size distribution. The model is inspired by landslide field data, but the principles may be applied to any system that experiences stabilizing failures, possesses a critical point, and is subject to an ongoing process of destabilization that includes occasional dramatic destabilizing events.

Crossover behavior in driven cascades.

We propose a model which explains how power-law crossover behavior can arise in a system which is capable of experiencing cascading failure. In our model the susceptibility of the system to cascades is described by a single number, the propagation power, which measures the ease with which cascades propagate. Physically, such a number could represent the density of unstable material in a system, its internal connectivity, or the mean susceptibility of its component parts to failure. We assume that the propagation power follows an upward drifting Brownian motion between cascades, and drops discontinuously each time a cascade occurs. Cascades are described by a continuous state branching process with distributional properties determined by the value of the propagation power when they occur. In common with many cascading models, pure power-law behavior is exhibited at a critical level of propagation power, and the mean cascade size diverges. This divergence constrains large systems to the subcritical region. We show that as a result, crossover behavior appears in the cascade distribution when an average is performed over the distribution of propagation power. We are able to analytically determine the exponents before and after the crossover.