The information content of symbolic sequences (such as nucleic- or amino acid sequences, but also neuronal firings or strings of letters) can be calculated from an ensemble of such sequences, but because information cannot be assigned to single seque
nces, we cannot correlate information to other observables attached to the sequence. Here we show that an information score obtained from multivariate (multiple-variable) correlations within sequences of a training ensemble can be used to predict observables of out-of-sample sequences with an accuracy that scales with the complexity of correlations, showing that functional information emerges from a hierarchy of multi-variable correlations.
Proteins perform critical processes in all living systems: converting solar energy into chemical energy, replicating DNA, as the basis of highly performant materials, sensing and much more. While an incredible range of functionality has been sampled
in nature, it accounts for a tiny fraction of the possible protein universe. If we could tap into this pool of unexplored protein structures, we could search for novel proteins with useful properties that we could apply to tackle the environmental and medical challenges facing humanity. This is the purpose of protein design. Sequence design is an important aspect of protein design, and many successful methods to do this have been developed. Recently, deep-learning methods that frame it as a classification problem have emerged as a powerful approach. Beyond their reported improvement in performance, their primary advantage over physics-based methods is that the computational burden is shifted from the user to the developers, thereby increasing accessibility to the design method. Despite this trend, the tools for assessment and comparison of such models remain quite generic. The goal of this paper is to both address the timely problem of evaluation and to shine a spotlight, within the Machine Learning community, on specific assessment criteria that will accelerate impact. We present a carefully curated benchmark set of proteins and propose a number of standard tests to assess the performance of deep learning based methods. Our robust benchmark provides biological insight into the behaviour of design methods, which is essential for evaluating their performance and utility. We compare five existing models with two novel models for sequence prediction. Finally, we test the designs produced by these models with AlphaFold2, a state-of-the-art structure-prediction algorithm, to determine if they are likely to fold into the intended 3D shapes.
A direct measurement of muscle and joint forces during typical human movements is desirable, e.g. to assess the effect of pain on these forces, and reduce joint forces to prevent further injury. For ethical and medical reasons, invasive joint force m
easurements are problematic, but computational models might provide a solution by predicting these forces. Since any modeling is an approximation, it is not yet clear how accurate predicted joint load forces and torques are for real-life biological movements. In contrast to real joints, it is, however possible to measure forces in implanted prostheses, providing an alternative method of validating the modelling approach. Therefore, the aim of this study was to investigate the accuracy of predicted forces in a knee joint during walking and squatting based on a computational musculoskeletal model, by comparing the model predictions with the corresponding real-life data gained from an instrumented knee prosthesis. Using calculated root mean squared error between the predicted and measured knee contact-forces, we found that musculoskeletal models can accurately predict knee joint forces. Furthermore, we demonstrated that the muscular coordination highly influences knee joint forces, as the knee joint forces were systematically reduced based on neuromuscular activation by -44% in walking and -15% in squatting. Our findings indicate that patients with a knee prosthesis may adapt their neuromuscular activation pattern to reduce joint forces during locomotion or everyday movements.
As a means to extract biomarkers from medical imaging, radiomics has attracted increased attention from researchers. However, reproducibility and performance of radiomics in low dose CT scans are still poor, mostly due to noise. Deep learning generat
ive models can be used to denoise these images and in turn improve radiomics reproducibility and performance. However, most generative models are trained on paired data, which can be difficult or impossible to collect. In this article, we investigate the possibility of denoising low dose CTs using cycle generative adversarial networks (GANs) to improve radiomics reproducibility and performance based on unpaired datasets. Two cycle GANs were trained: 1) from paired data, by simulating low dose CTs (i.e., introducing noise) from high dose CTs; and 2) from unpaired real low dose CTs. To accelerate convergence, during GAN training, a slice-paired training strategy was introduced. The trained GANs were applied to three scenarios: 1) improving radiomics reproducibility in simulated low dose CT images and 2) same-day repeat low dose CTs (RIDER dataset) and 3) improving radiomics performance in survival prediction. Cycle GAN results were compared with a conditional GAN (CGAN) and an encoder-decoder network (EDN) trained on simulated paired data.The cycle GAN trained on simulated data improved concordance correlation coefficients (CCC) of radiomic features from 0.87 to 0.93 on simulated noise CT and from 0.89 to 0.92 on RIDER dataset, as well improving the AUC of survival prediction from 0.52 to 0.59. The cycle GAN trained on real data increased the CCCs of features in RIDER to 0.95 and the AUC of survival prediction to 0.58. The results show that cycle GANs trained on both simulated and real data can improve radiomics reproducibility and performance in low dose CT and achieve similar results compared to CGANs and EDNs.
How dynamic interactions between nervous system regions in mammals performs online motor control remains an unsolved problem. Here we present a new approach using a minimal model comprising spinal cord, sensory and motor cortex, coupled by long conne
ctions that are plastic. It succeeds in learning how to perform reaching movements of a planar arm with 6 muscles in several directions from scratch. The model satisfies biological plausibility constraints, like neural implementation, transmission delays, local synaptic learning and continuous online learning. The model can go from motor babbling to reaching arbitrary targets in less than 10 minutes. However, because there is no cerebellum the movements are ataxic. As emergent properties, neural populations in motor cortex show directional tuning and oscillatory dynamics, and the spinal cord creates convergent force fields that add linearly. The model is extensible and may eventually lead to complete motor control simulation.
Bayesian inference is a popular and widely-used approach to infer phylogenies (evolutionary trees). However, despite decades of widespread application, it remains difficult to judge how well a given Bayesian Markov chain Monte Carlo (MCMC) run explor
es the space of phylogenetic trees. In this paper, we investigate the Monte Carlo error of phylogenies, focusing on high-dimensional summaries of the posterior distribution, including variability in estimated edge/branch (known in phylogenetics as split) probabilities and tree probabilities, and variability in the estimated summary tree. Specifically, we ask if there is any measure of effective sample size (ESS) applicable to phylogenetic trees which is capable of capturing the Monte Carlo error of these three summary measures. We find that there are some ESS measures capable of capturing the error inherent in using MCMC samples to approximate the posterior distributions on phylogenies. We term these tree ESS measures, and identify a set of three which are useful in practice for assessing the Monte Carlo error. Lastly, we present visualization tools that can improve comparisons between multiple independent MCMC runs by accounting for the Monte Carlo error present in each chain. Our results indicate that common post-MCMC workflows are insufficient to capture the inherent Monte Carlo error of the tree, and highlight the need for both within-chain mixing and between-chain convergence assessments.
We consider a population constituted by two types of individuals; each of them can produce offspring in two different islands (as a particular case the islands can be interpreted as active or dormant individuals). We model the evolution of the popula
tion of each type using a two-type Feller diffusion with immigration, and we study the frequency of one of the types, in each island, when the total population size in each island is forced to be constant at a dense set of times. This leads to the solution of a SDE which we call the asymmetric two-island frequency process. We derive properties of this process and obtain a large population limit when the total size of each island tends to infinity. Additionally, we compute the fluctuations of the process around its deterministic limit. We establish conditions under which the asymmetric two-island frequency process has a moment dual. The dual is a continuous-time two-dimensional Markov chain that can be interpreted in terms of mutation, branching, pairwise branching, coalescence, and a novel mixed selection-migration term. Also, we conduct a stability analysis of the limiting deterministic dynamical system and present some numerical results to study fixation and a new form of balancing selection. When restricting to the seedbank model, we observe that some combinations of the parameters lead to balancing selection. Besides finding yet another way in which genetic reservoirs increase the genetic variability, we find that if a population that sustains a seedbank competes with one that does not, the seed producers will have a selective advantage if they reproduce faster, but will not have a selective disadvantage if they reproduce slower: their worst case scenario is balancing selection.
Ability of smooth muscles to contract in response to distension plays a crucial role in motor function of intestine. Qualitative analysis of dynamical models using myogenic active property of smooth muscles has shown well agreement with physiologic d
ata. Considered as a self-regulatory unit, function of gastrointestinal (GI) segment is assumed to be regulated by integration of basis patterns providing accumulation and propagation of intestinal content. By implementing external, depending on neural system, variable to the previous model, and considering two attaches to one another reservoirs as a physical analogue of the segmental partition of intestine, a system of six ODE equations, three for each reservoir, describes coordinated wall motions and propagation of the content from one reservoir to another. It was shown that besides negative feedback (NFB), other functional patterns, namely positive feedback (PFB) and reciprocal links (RL) are involved in regulations of filling-emptying cycle. Being integrated in a whole functional system these three patterns expressed in a matrix form represent basis elements of imaginary part of coquaternion which with identity basis component is an algebraically closed structure under addition and multiplication of its elements. A coquaternion ring may be considered as a model of inner self-regulatory functional structure providing coordinated wall motions of GI tract portions.
Arid zones contain a diverse set of microbes capable of survival under dry conditions, some of which can form relationships with plants under drought stress conditions to improve plant health. We studied squash (Cucurbita pepo L.) root microbiome und
er historically arid and humid sites, both in situ and performing a common garden experiment. Plants were grown in soils from sites with different drought levels, using in situ collected soils as the microbial source. We described and analyzed bacterial diversity by 16S rRNA gene sequencing (N=48) from the soil, rhizosphere, and endosphere. Proteobacteria were the most abundant phylum present in humid and arid samples, while Actinobacteriota abundance was higher in arid ones. The Beta-diversity analyses showed split microbiomes between arid and humid microbiomes, and aridity and soil pH levels could explain it. These differences between humid and arid microbiomes were maintained in the common garden experiment, showing that it is possible to transplant in situ diversity to the greenhouse. We detected a total of 1009 bacterial genera; 199 exclusively associated with roots under arid conditions. With shotgun metagenomic sequencing of rhizospheres (N=6), we identified 2969 protein families in the squash core metagenome and found an increased number of exclusively protein families from arid (924) than humid samples (158). We found arid conditions enriched genes involved in protein degradation and folding, oxidative stress, compatible solute synthesis, and ion pumps associated with osmotic regulation. Plant phenotyping allowed us to correlate bacterial communities with plant growth. Our study revealed that it is possible to evaluate microbiome diversity ex-situ and identify critical species and genes involved in plant-microbe interactions in historically arid locations.
Noise in spiking neurons is commonly modeled by a noisy input current or by generating output spikes stochastically with a voltage-dependent hazard rate (escape noise). While input noise lends itself to modeling biophysical noise processes, the pheno
menological escape noise is mathematically more tractable. Using the level-crossing theory for differentiable Gaussian processes, we derive an approximate mapping between colored input noise and escape noise in leaky integrate-and-fire neurons. This mapping requires the first-passage-time (FPT) density of an overdamped Brownian particle driven by colored noise with respect to an arbitrarily moving boundary. Starting from the Wiener-Rice series for the FPT density, we apply the second-order decoupling approximation of Stratonovich to the case of moving boundaries and derive a simplified hazard-rate representation that is local in time and numerically efficient. This simplification requires the calculation of the non-stationary auto-correlation function of the level-crossing process: For exponentially correlated input noise (Ornstein-Uhlenbeck process), we obtain an exact formula for the zero-lag auto-correlation as a function of noise parameters, mean membrane potential and its speed, as well as an exponential approximation of the full auto-correlation function. The theory well predicts the FPT and interspike interval densities as well as the population activities obtained from simulations with time-dependent stimulus or boundary. The agreement with simulations is strongly enhanced compared to a first-order decoupling approximation that neglects correlations between level crossings. The second-order approximation also improves upon a previously proposed theory in the subthreshold regime. Depending on a simplicity-accuracy trade-off, all considered approximations represent useful mappings from colored input noise to escape noise.
