DeepQMC: An open-source software suite for variational optimization of deep-learning molecular wave functions.

Computing accurate yet efficient approximations to the solutions of the electronic Schrödinger equation has been a paramount challenge of computational chemistry for decades. Quantum Monte Carlo methods are a promising avenue of development as their core algorithm exhibits a number of favorable properties: it is highly parallel and scales favorably with the considered system size, with an accuracy that is limited only by the choice of the wave function Ansatz. The recently introduced machine-learned parametrizations of quantum Monte Carlo Ansätze rely on the efficiency of neural networks as universal function approximators to achieve state of the art accuracy on a variety of molecular systems. With interest in the field growing rapidly, there is a clear need for easy to use, modular, and extendable software libraries facilitating the development and adoption of this new class of methods. In this contribution, the DeepQMC program package is introduced, in an attempt to provide a common framework for future investigations by unifying many of the currently available deep-learning quantum Monte Carlo architectures. Furthermore, the manuscript provides a brief introduction to the methodology of variational quantum Monte Carlo in real space, highlights some technical challenges of optimizing neural network wave functions, and presents example black-box applications of the program package. We thereby intend to make this novel field accessible to a broader class of practitioners from both the quantum chemistry and the machine learning communities.

Electronic excited states in deep variational Monte Carlo.

Obtaining accurate ground and low-lying excited states of electronic systems is crucial in a multitude of important applications. One ab initio method for solving the Schrödinger equation that scales favorably for large systems is variational quantum Monte Carlo (QMC). The recently introduced deep QMC approach uses ansatzes represented by deep neural networks and generates nearly exact ground-state solutions for molecules containing up to a few dozen electrons, with the potential to scale to much larger systems where other highly accurate methods are not feasible. In this paper, we extend one such ansatz (PauliNet) to compute electronic excited states. We demonstrate our method on various small atoms and molecules and consistently achieve high accuracy for low-lying states. To highlight the method's potential, we compute the first excited state of the much larger benzene molecule, as well as the conical intersection of ethylene, with PauliNet matching results of more expensive high-level methods.

Convergence to the fixed-node limit in deep variational Monte Carlo.

Variational quantum Monte Carlo (QMC) is an ab initio method for solving the electronic Schrödinger equation that is exact in principle, but limited by the flexibility of the available Ansätze in practice. The recently introduced deep QMC approach, specifically two deep-neural-network Ansätze PauliNet and FermiNet, allows variational QMC to reach the accuracy of diffusion QMC, but little is understood about the convergence behavior of such Ansätze. Here, we analyze how deep variational QMC approaches the fixed-node limit with increasing network size. First, we demonstrate that a deep neural network can overcome the limitations of a small basis set and reach the mean-field (MF) complete-basis-set limit. Moving to electron correlation, we then perform an extensive hyperparameter scan of a deep Jastrow factor for LiH and H4 and find that variational energies at the fixed-node limit can be obtained with a sufficiently large network. Finally, we benchmark MF and many-body Ansätze on H2O, increasing the fraction of recovered fixed-node correlation energy of single-determinant Slater-Jastrow-type Ansätze by half an order of magnitude compared to previous variational QMC results, and demonstrate that a single-determinant Slater-Jastrow-backflow version of the Ansatz overcomes the fixed-node limitations. This analysis helps understand the superb accuracy of deep variational Ansätze in comparison to the traditional trial wavefunctions at the respective level of theory and will guide future improvements of the neural-network architectures in deep QMC.

Voluntary and electrically-induced muscle fatigue differently affect postural control mechanisms in unipedal stance.

The repetition of muscle contractions is likely to generate fatigue which can provoke alterations of postural control. Regulatory mechanisms can be triggered to counteract these alterations. However, these mechanisms would occur only when fatigue is induced through voluntary (VOL) contractions and not with electrically stimulated (ES) contractions. Hence the aim was to compare the effects of VOL and ES fatiguing contractions inducing a similar level of strength loss on unipedal postural control (assessed by means of force platform and EMG measurements), maximal voluntary contraction (MVC) and central activation ratio (CAR) to characterize the alterations induced by both modalities of fatigue and the associated regulatory mechanisms. Results showed that the VOL exercise induced a significant decrease of the CAR whereas the ES exercise did not, thus illustrating that central fatigue was present only after voluntary contractions. The VOL exercise also induced greater postural disturbances and larger regulatory mechanisms than the ES exercise, which also induced postural regulatory mechanisms. The present study reveals that postural control mechanisms are modulated according to the nature of the fatiguing contractions, likely due integration of specific fatigue signals according to the modality of the contraction. Because of a larger neurophysiological impact of VOL than ES fatiguing contractions due to greater central disturbances, VOL exercise-induced larger regulatory mechanisms. Nevertheless, the presence of regulatory mechanisms with ES contractions clearly underlines the ability of the central nervous system to display an accurate motor control following acute externally induced neuromuscular perturbations.

Collective hydrogen-bond rearrangement dynamics in liquid water.

We investigate barrier-crossing processes corresponding to collective hydrogen-bond rearrangements in liquid water using Markov state modeling techniques. The analysis is based on trajectories from classical molecular dynamics simulations and accounts for the full dynamics of relative angular and separation coordinates of water clusters and requires no predefined hydrogen bond criterium. We account for the complete 12-dimensional conformational subspace of three water molecules and distinguish five well-separated slow dynamic processes with relaxation times in the picosecond range, followed by a quasi-continuum spectrum of faster modes. By analysis of the Markov eigenstates, these processes are shown to correspond to different collective interchanges of hydrogen-bond donors and acceptors. Using a projection onto hydrogen-bond states, we also analyze the switching of one hydrogen bond between two acceptor water molecules and derive the complete transition network. The most probable pathway corresponds to a direct switch without an intermediate, in agreement with previous studies. However, a considerable fraction of paths proceeds along alternative routes that involve different intermediate states with short-lived alternative hydrogen bonds or weakly bound states.

Detection of Coxiella burnetii DNA in Peridomestic and Wild Animals and Ticks in an Endemic Region (Canary Islands, Spain).

Coxiella burnetii, the etiological agent of human Q fever, can infect mammals, birds, and arthropods. The Canary Islands (Spain) are considered an endemic territory, with a high prevalence in both humans and livestock. Nonetheless, there is no epidemiological information about the wild and peridomestic cycles of C. burnetii. Tissue samples from rodents on farms (100) and wild rabbits (129) were collected and assessed by PCR to detect C. burnetii DNA. In parallel, ticks were also collected from vegetation (1169), livestock (335), domestic dogs (169), and wild animals (65). Globally, eight rodents (8%) and two rabbits (1.5%) were found to be positive, with the spleen being the most affected organ. Tick species identified were Hyalomma lusitanicum, Rhipicephalus turanicus, Rhipicephalus sanguineus, and Rhipicephalus pusillus. Hyalomma lusitanicum (80%) was the main species identified in vegetation, livestock, and wild animals, whereas Rhipicephalus sanguineus was the most prevalent in domestic dogs. Overall, C. burnetii DNA was detected in 6.1% of the processed ticks, distributed between those removed from livestock (11.3%), domestic dogs (6.9%), and from wild animals (6%). Ticks from vegetation were all negative. Results suggest that, in the Canary Islands, C. burnetii develops in a peridomestic rather than a wild cycle.

Molecular dynamics simulations data of the twenty encoded amino acids in different force fields.

We present extensive all-atom Molecular Dynamics (MD) simulation data of the twenty encoded amino acids in explicit water, simulated with different force fields. The termini of the amino acids have been capped to ensure that the dynamics of the Φ and ψ torsion angles are analogues to the dynamics within a peptide chain. We use representatives of each of the four major force field families: AMBER ff-99SBILDN [1], AMBER ff-03 [2], OPLS-AA/L [3], CHARMM27 [4] and GROMOS43a1 [5], [6]. Our data represents a library and test bed for method development for MD simulations and for force fields development. Part of the data set has been previously used for comparison of the dynamic properties of force fields (Vitalini et al., 2015) [7] and for the construction of peptide basis functions for the variational approach to molecular kinetics [8].

A Basis Set for Peptides for the Variational Approach to Conformational Kinetics.

Although Markov state models have proven to be powerful tools in resolving the complex features of biomolecular kinetics, the discretization of the conformational space has been a bottleneck since the advent of the method. A recently introduced variational approach, which uses basis functions instead of crisp conformational states, opened up a route to construct kinetic models in which the discretization error can be controlled systematically. Here, we develop and test a basis set for peptides to be used in the variational approach. The basis set is constructed by combining local residue-centered kinetic modes that are obtained from kinetic models of terminally blocked amino acids. Using this basis set, we model the conformational kinetics of two hexapeptides with sequences VGLAPG and VGVAPG. Six basis functions are sufficient to represent the slow kinetic modes of these peptides. The basis set also allows for a direct interpretation of the slow kinetic modes without an additional clustering in the space of the dominant eigenvectors. Moreover, changes in the conformational kinetics due to the exchange of leucine in VGLAPG to valine in VGVAPG can be directly quantified by comparing histograms of the basis set expansion coefficients.

Dynamic properties of force fields.

Molecular-dynamics simulations are increasingly used to study dynamic properties of biological systems. With this development, the ability of force fields to successfully predict relaxation timescales and the associated conformational exchange processes moves into focus. We assess to what extent the dynamic properties of model peptides (Ac-A-NHMe, Ac-V-NHMe, AVAVA, A10) differ when simulated with different force fields (AMBER ff99SB-ILDN, AMBER ff03, OPLS-AA/L, CHARMM27, and GROMOS43a1). The dynamic properties are extracted using Markov state models. For single-residue models (Ac-A-NHMe, Ac-V-NHMe), the slow conformational exchange processes are similar in all force fields, but the associated relaxation timescales differ by up to an order of magnitude. For the peptide systems, not only the relaxation timescales, but also the conformational exchange processes differ considerably across force fields. This finding calls the significance of dynamic interpretations of molecular-dynamics simulations into question.

A transversal study on antibodies against selected pathogens in dromedary camels in the Canary Islands, Spain.

The Canary Islands contain the most important dromedary camel (Camelus dromedarius) population in the European Union and are the main export point of dromedaries to continental Europe and Latin America. We investigated the presence of antibodies against relevant disease agents in 100 Canarian camel sera. Selected blood samples of the same animals were also tested by PCR. Sera were tested for antibodies against Bluetongue virus (BTV; 0%), Bovine Viral Diarrhoea virus (BVDV; 0%), Camelpox virus (CPV; 8% by serum neutralization, 16% by ELISA), Peste des Petits Ruminants virus (PPRV, 0%), Rift Valley Fever virus (RVFV; 0%) and West Nile Fever virus (WNV; 3%), the bacterial pathogens Anaplasma sp. (3%), Brucella sp. (1%), Coxiella burnetii (19%), Mycobacterium avium paratuberculosis (MAP; 22%), Mycobacterium tuberculosis complex (MTC; 10%) and Rickettsia sp. (83%), and the parasites Toxoplasma gondii (36%) and Neospora caninum (86%). The most remarkable findings were the detection of antibodies against CPV and the high antibody prevalence against C. burnetii, Rickettsia sp., T. gondii and N. caninum. By PCR, we found no C. burnetii, N. caninum and Anaplasma sp. DNA in the tested samples. However, Rickettsia sp. DNA was detected in six antibody positive tested samples. These results should be taken into consideration in order to implement adequate control measures and avoid a potential dissemination of infections to other territories.

Pharmacological blockade of IL-1ß/IL-1 receptor type 1 axis during epileptogenesis provides neuroprotection in two rat models of temporal lobe epilepsy.

We studied whether pharmacological blockade of the IL-1ß-mediated signaling, rapidly activated in forebrain by epileptogenic injuries, affords neuroprotection in two different rat models of status epilepticus (SE). As secondary outcome, we measured treatment's effect on SE-induced epileptogenesis. IL-1ß signaling was blocked by systemic administration of two antiinflammatory drugs, namely human recombinant IL-1 receptor antagonist (anakinra), the naturally occurring and clinically used competitive IL-1 receptor type 1 antagonist, and VX-765 a specific non-peptide inhibitor of IL-1ß cleavage and release. Antiinflammatory drugs were given 60min after antiepileptic (AED) drug-controlled SE induced by pilocarpine, or 180min after unrestrained electrical SE, for 7days using a protocol yielding therapeutic drug levels in brain. This drug combination significantly decreased both IL-1ß expression in astrocytes and cell loss in rat forebrain. Neuroprotection and the antiinflammatory effect were more pronounced in the electrical SE model. Onset of epilepsy, and frequency and duration of seizures 3months after electrical SE were not significantly modified. Transcriptomic analysis in the hippocampus showed that the combined treatment did not affect the broad inflammatory response induced by SE during epileptogenesis. In particular, the treatment did not prevent the induction of the complement system and Toll-like receptors, both contributing to cell loss and seizure generation. We conclude that the IL-1ß signaling represents an important target for reducing cell loss after SE. The data highlight a new class of clinically tested agents affording neuroprotection after a delayed post-injury intervention. Earlier blockade of this rapid onset inflammatory pathway during SE, or concomitant treatment with antiinflammatory drugs targeting additional components of the broad inflammatory response to SE, or co-treatment with AEDs, is likely to be required for optimizing beneficial outcomes.

Molecular method for the characterization of Coxiella burnetii from clinical and environmental samples: variability of genotypes in Spain.

BACKGROUND: Coxiella burnetii is a highly clonal microorganism which is difficult to culture, requiring BSL3 conditions for its propagation. This leads to a scarce availability of isolates worldwide. On the other hand, published methods of characterization have delineated up to 8 different genomic groups and 36 genotypes. However, all these methodologies, with the exception of one that exhibited limited discriminatory power (3 genotypes), rely on performing between 10 and 20 PCR amplifications or sequencing long fragments of DNA, which make their direct application to clinical samples impracticable and leads to a scarce accessibility of data on the circulation of C. burnetii genotypes. RESULTS: To assess the variability of this organism in Spain, we have developed a novel method that consists of a multiplex (8 targets) PCR and hybridization with specific probes that reproduce the previous classification of this organism into 8 genomic groups, and up to 16 genotypes. It allows for a direct characterization from clinical and environmental samples in a single run, which will help in the study of the different genotypes circulating in wild and domestic cycles as well as from sporadic human cases and outbreaks. The method has been validated with reference isolates. A high variability of C. burnetii has been found in Spain among 90 samples tested, detecting 10 different genotypes, being those adaA negative associated with acute Q fever cases presenting as fever of intermediate duration with liver involvement and with chronic cases. Genotypes infecting humans are also found in sheep, goats, rats, wild boar and ticks, and the only genotype found in cattle has never been found among our clinical samples. CONCLUSIONS: This newly developed methodology has permitted to demonstrate that C. burnetii is highly variable in Spain. With the data presented here, cattle seem not to participate in the transmission of C. burnetii to humans in the samples studied, while sheep, goats, wild boar, rats and ticks share genotypes with the human population.

Anticonvulsant effects and behavioural outcomes of rAAV serotype 1 vector-mediated neuropeptide Y overexpression in rat hippocampus.

Neuropeptide Y (NPY) is an endogenous peptide with powerful anticonvulsant properties. Its overexpression in the rat hippocampus, mediated by the local application of recombinant adeno-associated viral (rAAV) vectors carrying the human NPY gene, results in significant reduction of seizures in acute and chronic seizure models. In this study, we characterized a more efficient rAAV-NPY vector to improve cell transfection in the injected area. The changes included pseudotyping with the AAV vector serotype 1 (rAAV1), and using the strong constitutive hybrid CBA promoter, which contains a cytomegalovirus enhancer and chicken beta-actin promoter sequences. We compared NPY expression and the associated anticonvulsant effects of this new vector, with those mediated by the former rAAV vector with chimeric serotype 1/2 (rAAV1/2). In addition, we investigated whether rAAV serotype 1 vector-mediated chronic NPY overexpression causes behavioural deficits that may detract from the clinical utility of this therapeutic approach. We report that rAAV-NPY serotype 1 vector has significantly improved anticonvulsant activity when compared with serotype 1/2 vector, as assessed by measuring EEG seizure activity in kainic acid treated rats. rAAV1-mediated NPY overexpression in naive rats did not result in alterations of physiological functions such as learning and memory, anxiety and locomotor activity. In addition, we did not observe glia activation, or humoral immune responses against serotype 1 vector, which could inactivate gene expression. Our findings show that rAAV1-NPY vector with the CBA promoter mediates powerful anticonvulsant effects and seems to be safe in rodents, thus it may be considered a vector of choice for possible clinical applications.

Acute induction of epileptiform discharges by pilocarpine in the in vitro isolated guinea-pig brain requires enhancement of blood-brain barrier permeability.

Systemic application of the muscarinic agonist, pilocarpine, is commonly utilized to induce an acute status epilepticus that evolves into a chronic epileptic condition characterized by spontaneous seizures. Recent findings suggest that the status epilepticus induced by pilocarpine may be triggered by changes in the blood-brain barrier (BBB) permeability. We tested the role of the BBB in an acute pilocarpine model by using the in vitro model brain preparation and compared our finding with in vivo data. Arterial perfusion of the in vitro isolated guinea-pig brain with <1 mM pilocarpine did not cause epileptiform activity, but rather reduced synaptic transmission and induced steady fast (20-25 Hz) oscillatory activity in limbic cortices. These effects were reversibly blocked by co-perfusion of the muscarinic antagonist atropine sulfate (5 microM). Brain pilocarpine measurements in vivo and in vitro suggested modest BBB penetration. Pilocarpine induced epileptiform discharges only when perfused with compounds that enhance BBB permeability, such as bradykinin (n=2) or histamine (n=10). This pro-epileptic effect was abolished when the BBB-impermeable muscarinic antagonist atropine methyl bromide (5 microM) was co-perfused with histamine and pilocarpine. In the absence of BBB permeability enhancing drugs, pilocarpine induced epileptiform activity only after arterial perfusion at concentrations >10 mM. Ictal discharges correlated with a high intracerebral pilocarpine concentration measured by high pressure liquid chromatography. We propose that acute epileptiform discharges induced by pilocarpine treatment in the in vitro isolated brain preparation are mediated by a dose-dependent, atropine-sensitive muscarinic effect promoted by an increase in BBB permeability. Pilocarpine accumulation secondary to BBB permeability changes may contribute to in vivo ictogenesis in the pilocarpine epilepsy model.

Effect of expertise and visual contribution on postural control in soccer.

This work analyzes the postural performance and the use of visual information in soccer players according to their level of competition. Two groups of healthy soccer players were investigated at the mid-competition season: an amateur (AM) (n=15) group composed of regional-level players and a professional group at a national level (PRO) (n=15). Posture was assessed by measuring the center of foot pressure (COP) with a force platform during a test (51.2 s) of bipedal quiet standing posture. The test was completed with eyes open (the subjects looked at a fixed-level target at a distance of 2 m) and closed (they kept their gaze in a straight-ahead direction). The statistical analysis showed that PRO soccer players were more stable than AM soccer players. Moreover, the contribution of vision in postural maintenance was less important in the PRO players than in the AM players. The present study suggests that intense training allows PRO soccer players to become less dependent on vision to control their posture such that vision can be dedicated to treating the information, that emanates from the game.

Modifications of anticipatory postural adjustments in a rock climbing task: the effect of supporting wall inclination.

The aim of this study was to analyse the influence of initial postural constraint on the realisation of a leg release in a rock climbing task. Two conditions were tested: a vertical posture and an overhanging posture. The overhanging posture was characterised by a large sustentation base, which enhanced the mechanical possibilities of the system. Subjects had to release their right foot in both postural conditions. In the vertical posture, movement's effectuation was associated with anticipatory postural adjustments (APAs). In the overhanging posture, the movement was performed without APAs. The results indicated that APAs were modulated according to the possibilities of force creation of the system. Hence, the disappearance of APAs in the overhanging posture was explained by the efficiency of the system to create the impulse necessary to perform the task.

Is postural control affected by expertise in alpine skiing?

OBJECTIVES: This study examined the postural performance of two groups of male skiers competing at different levels and the consequences on postural control of the suppression of visual afferences by eye closure. METHODS: Seven national level (NAT) skiers and 7 regional level (REG) skiers were asked to stand as still as possible on a force platform with eyes opened and closed and while wearing or not wearing their ski boots in a stable posture and in two unstable postures (in the sagittal or frontal plane). Postural performance was assessed with centre of foot pressure measurements. RESULTS: REG and NAT skiers were similarly influenced by the absence of visual information and presented similar postural performance when tests were performed with ski boots. However, without ski boots, REG skiers displayed better postural performance than NAT skiers. CONCLUSIONS: The inferior postural performance of NAT skiers without ski boots could be a long term effect of repetitive wearing of ski boots, which impairs postural performance by restricting the range of motion of the ankle-foot complex. Since individuals with decreased postural performance are believed to be more susceptible to ankle injury than those with finer postural control, NAT skiers should benefit from specific training aimed at improving postural ability and preventing ankle injury.

[Neuromuscular effects of superimposed and combined transcutaneous electrical stimulation with voluntary activity: a review]. / Effets neuromusculaires de l'électrostimulation transcutanée surimposée et combinée à l'activité volontaire: une revue.

With voluntary muscular contraction (VOL), small motor units (MUs) are recruited before large MUs are (a submaximal muscular contraction recruits only small MUs), whereas electrical stimulation (ES) tends to reverse the recruitment order. On the basis of this observation, some authors have tested the physiological effects of ES superimposed simultaneously with VOL (superimposed technique [ST]) or separately (combined technique [CT]). With healthy subjects, ST does not recruit more MUs than VOL, except with eccentric contractions. After health subjects undergo training programs, ST appears to be as efficient as VOL in enhancing subjects' neuromuscular qualities. Nevertheless, the use of CT seems more effective than VOL. In postsurgical rehabilitation, both ST and CT are more effective than VOL. Actually, following knee surgery, ST and CT compensate for volume and muscle strength deficits with more efficiency than does VOL.

Influence of steep gradient supporting walls in rock climbing: biomechanical analysis.

This study analyses the reaction forces and variations of rock climbing in vertical and overhanging positions. Subjects voluntarily released their right foot and regained equilibrium. In the overhanging position the quadrupedal state was characterised by a significant involvement of the arms to prevent fall. Moreover, the horizontal forces applied to the holds were less important, which suggests that equilibrium was easier to maintain than in the vertical position. The tripedal state was characterised by less extensive contralateral supporting force transfer on the remaining holds in the overhanging position, which reinforces safety.

Single-breath measurements of pulmonary oxygen uptake and gas flow rates for ventilator management in ARDS.

Monitoring data in critical care and anesthesiology should be displayed to present a rapid and easily comprehensible definition of the patient's clinical status. A graphic computer display of the analog output of gas flow rates and the O(2) and CO(2) concentrations of respiratory gases profiles the expired breath for an estimation of pulmonary function and gas exchange. An estimate of pulmonary perfusion, cardiac output, and the general adequacy of cardiovascular circulation is obtained from the computer calculation of O(2) uptake and CO(2) elimination, dead space, and alveolar ventilation. Adjunctive data from the spirometric measurements of airway pressures, volumes, and compliance, supplemented by hemodynamic monitoring, aids in the diagnosis of physiologic changes. For > 10 years, we have used this system to monitor patients who are anesthetized, sedated, and receiving mechanical ventilation during anesthesia and surgery, and recently have extended the technique to intensive care areas. Our experience has shown good correlation of changes in the computer-assisted expired breath analysis with coinciding clinical events, including upper airway obstruction, bronchospasm, and alveolar volume/pulmonary capillary blood flow impairment. To demonstrate the use of this system, we describe the ventilator management for a patient with severe ARDS. In this patient, changes in ventilator management, including pressure control ventilation, improved pulmonary O(2) uptake (mean, 18.7 vs 8.5 mL/breath), CO(2) elimination (mean, 17 vs 13 mL/breath), and compliance (mean, 29.7 vs 19.0 mL/cm H(2)O), were compared with intermittent mandatory ventilation.