ELIXIR and Toxicology: a community in development.

Toxicology has been an active research field for many decades, with academic, industrial and government involvement. Modern omics and computational approaches are changing the field, from merely disease-specific observational models into target-specific predictive models. Traditionally, toxicology has strong links with other fields such as biology, chemistry, pharmacology and medicine. With the rise of synthetic and new engineered materials, alongside ongoing prioritisation needs in chemical risk assessment for existing chemicals, early predictive evaluations are becoming of utmost importance to both scientific and regulatory purposes. ELIXIR is an intergovernmental organisation that brings together life science resources from across Europe. To coordinate the linkage of various life science efforts around modern predictive toxicology, the establishment of a new ELIXIR Community is seen as instrumental. In the past few years, joint efforts, building on incidental overlap, have been piloted in the context of ELIXIR. For example, the EU-ToxRisk, diXa, HeCaToS, transQST, and the nanotoxicology community have worked with the ELIXIR TeSS, Bioschemas, and Compute Platforms and activities. In 2018, a core group of interested parties wrote a proposal, outlining a sketch of what this new ELIXIR Toxicology Community would look like. A recent workshop (held September 30th to October 1st, 2020) extended this into an ELIXIR Toxicology roadmap and a shortlist of limited investment-high gain collaborations to give body to this new community. This Whitepaper outlines the results of these efforts and defines our vision of the ELIXIR Toxicology Community and how it complements other ELIXIR activities.

Introducing WikiPathways as a Data-Source to Support Adverse Outcome Pathways for Regulatory Risk Assessment of Chemicals and Nanomaterials.

A paradigm shift is taking place in risk assessment to replace animal models, reduce the number of economic resources, and refine the methodologies to test the growing number of chemicals and nanomaterials. Therefore, approaches such as transcriptomics, proteomics, and metabolomics have become valuable tools in toxicological research, and are finding their way into regulatory toxicity. One promising framework to bridge the gap between the molecular-level measurements and risk assessment is the concept of adverse outcome pathways (AOPs). These pathways comprise mechanistic knowledge and connect biological events from a molecular level toward an adverse effect outcome after exposure to a chemical. However, the implementation of omics-based approaches in the AOPs and their acceptance by the risk assessment community is still a challenge. Because the existing modules in the main repository for AOPs, the AOP Knowledge Base (AOP-KB), do not currently allow the integration of omics technologies, additional tools are required for omics-based data analysis and visualization. Here we show how WikiPathways can serve as a supportive tool to make omics data interoperable with the AOP-Wiki, part of the AOP-KB. Manual matching of key events (KEs) indicated that 67% could be linked with molecular pathways. Automatic connection through linkage of identifiers between the databases showed that only 30% of AOP-Wiki chemicals were found on WikiPathways. More loose linkage through gene names in KE and Key Event Relationships descriptions gave an overlap of 70 and 71%, respectively. This shows many opportunities to create more direct connections, for example with extended ontology annotations, improving its interoperability. This interoperability allows the needed integration of omics data linked to the molecular pathways with AOPs. A new AOP Portal on WikiPathways is presented to allow the community of AOP developers to collaborate and populate the molecular pathways that underlie the KEs of AOP-Wiki. We conclude that the integration of WikiPathways and AOP-Wiki will improve risk assessment because omics data will be linked directly to KEs and therefore allow the comprehensive understanding and description of AOPs. To make this assessment reproducible and valid, major changes are needed in both WikiPathways and AOP-Wiki.

How Adverse Outcome Pathways Can Aid the Development and Use of Computational Prediction Models for Regulatory Toxicology.

Efforts are underway to transform regulatory toxicology and chemical safety assessment from a largely empirical science based on direct observation of apical toxicity outcomes in whole organism toxicity tests to a predictive one in which outcomes and risk are inferred from accumulated mechanistic understanding. The adverse outcome pathway (AOP) framework provides a systematic approach for organizing knowledge that may support such inference. Likewise, computational models of biological systems at various scales provide another means and platform to integrate current biological understanding to facilitate inference and extrapolation. We argue that the systematic organization of knowledge into AOP frameworks can inform and help direct the design and development of computational prediction models that can further enhance the utility of mechanistic and in silico data for chemical safety assessment. This concept was explored as part of a workshop on AOP-Informed Predictive Modeling Approaches for Regulatory Toxicology held September 24-25, 2015. Examples of AOP-informed model development and its application to the assessment of chemicals for skin sensitization and multiple modes of endocrine disruption are provided. The role of problem formulation, not only as a critical phase of risk assessment, but also as guide for both AOP and complementary model development is described. Finally, a proposal for actively engaging the modeling community in AOP-informed computational model development is made. The contents serve as a vision for how AOPs can be leveraged to facilitate development of computational prediction models needed to support the next generation of chemical safety assessment.

Estimation of the error between experimental tetanic force curves of MUs of rat medial gastrocnemius muscle and their models by summation of equal successive contractions.

More accurate muscle models require appropriate modelling of individual twitches of motor units (MUs) and their unfused tetanic contractions. It was shown in our previous papers, using a few MUs, that modelling of unfused tetanic force curves by summation of equal twitches is not accurate, especially for slow MUs. The aim of this study was to evaluate this inaccuracy using a statistical number of MUs of the rat medial gastrocnemius muscle (15 of slow, 15 of fast resistant and 15 of fast fatigable type). Tetanic contractions were evoked by trains of 41 stimuli at random interpulse intervals and different mean frequencies, resembling discharge patterns observed during natural muscle activity. The tetanic curves were calculated by the summation of equal twitches according to the respective experimental patterns. The previously described 6-parameter analytical function for twitch modelling was used. Comparisons between the experimental and the modelled curves were made using two coefficients: the fit coefficient and the area coefficient. The errors between modelled and experimental tetanic forces were substantially different between the three MU types. The error was the most significant for slow MUs, which develop much higher forces in real contractions than could be predicted based on the summation of equal twitches, while the smallest error was observed for FF MUs--their recorded tetanic forces were similar to those predicted by modelling. The obtained results indicate the importance of the inclusion of the type-specific non-linearity in the summation of successive twitch-like contractions of MUs in order to increase the reliability of modelling skeletal muscle force.

Dynamic changes of twitchlike responses to successive stimuli studied by decomposition of motor unit tetanic contractions in rat medial gastrocnemius.

Unfused tetanic contractions evoked by trains of stimuli at variable interpulse intervals (IPIs) were recorded for 10 fast fatigable (FF), 10 fast resistant (FR), and 10 slow (S) motor units (MUs) and subsequently decomposed with a mathematical algorithm into trains of twitch-shape responses to successive stimuli. The mean stimulation frequencies were matched for each MU to evoke tetani of similar fusion degrees, whereas the variability range of IPIs was in each case 50-150% of the mean IPI. Force and time parameters of decomposed twitches were analyzed and related to the first response. Considerable variability of the analyzed twitch parameters was observed in each MU, although the largest range of variability occurred in slow MUs. In general, the decomposed twitch responses had longer duration and higher force than single-twitch contractions, although for nine FF and six FR MUs some of the decomposed responses were slightly weaker (but not faster) than the first twitches of these MUs. Comparison of the strongest decomposed twitch to the first decomposed twitch revealed ratios of forces up to 2.35, 3.33, and 6.89 for FF, FR, and S MUs and ratios of force-time areas up to 3.54, 4.67, and 14.26 for FF, FR, and S MUs, whereas for the contraction times the ratios of the longest decomposed twitch to the first twitch amounted to 2.46, 2.07, and 3.52 for FF, FR, and S MUs, respectively. The results indicate that contractile responses to successive action potentials are considerably variable, especially for slow MUs.

An approach for simulation of the muscle force modeling it by summation of motor unit contraction forces.

Muscle force is due to the cumulative effect of repetitively contracting motor units (MUs). To simulate the contribution of each MU to whole muscle force, an approach implemented in a novel computer program is proposed. The individual contraction of an MU (the twitch) is modeled by a 6-parameter analytical function previously proposed; the force of one MU is a sum of its contractions due to an applied stimulation pattern, and the muscle force is the sum of the active MUs. The number of MUs, the number of slow, fast-fatigue-resistant, and fast-fatigable MUs, and their six parameters as well as a file with stimulation patterns for each MU are inputs for the developed software. Different muscles and different firing patterns can be simulated changing the input data. The functionality of the program is illustrated with a model consisting of 30 MUs of rat medial gastrocnemius muscle. The twitches of these MUs were experimentally measured and modeled. The forces of the MUs and of the whole muscle were simulated using different stimulation patterns that included different regular, irregular, synchronous, and asynchronous firing patterns of MUs. The size principle of MUs for recruitment and derecruitment was also demonstrated using different stimulation paradigms.

Conformational coverage by a genetic algorithm: saturation of conformational space.

The molecular modeling is traditionally based on analysis of minimum energy conformers. Such simplifying assumptions could doom to failure the modeling studies given the significant variation of the geometric and electronic characteristics across the multitude of energetically reasonable conformers representing the molecules. Moreover, it has been found that the lowest energy conformers of chemicals are not necessarily the active ones with respect to various endpoints. Hence, the selection of active conformers appears to be as important as the selection of molecular descriptors in the modeling process. In this respect, we have developed effective tools for conformational analysis based on a genetic algorithm (GA), published in J. Chem. Inf. Comput. Sci. (1994, 34, 234; 1999, 39 (6), 997) and J. Chem. Inf. Model. (2005, 45 (2), 283). This paper presents a further improvement of the evolutionary algorithm for conformer generation minimizing the sensitivity of conformer distributions from the effect of smoothing parameter and improving the reproducibility of conformer distributions given the nondeterministic character of the genetic algorithm (GA). The ultimate goal of the saturation is to represent the conformational space of chemicals with an optimal number of conformers providing a stable conformational distribution which cannot be further perturbed by the addition of new conformers. The generation of stable conformational distributions of chemicals by a limited number of conformers will improve the adequacy of the subsequent molecular modeling analysis. The impact of the saturation procedure on conformer distributions in a specific structural space is illustrated by selected examples. The effect of the procedure on similarity assessment between chemicals is discussed.

Variability of the twitch parameters of the rat medial gastrocnemius motor units--experimental and modeling study.

In the present study a previously proposed model of a twitch based on an analytical function with four-parameters (lead, contraction and half-relaxation times and maximum force of the twitch) was validated on 115 motor units (MUs), divided into slow (S), fast-fatigue resistant (FR) and fast fatigable (FF) types. The original records were collected from electrophysiological experiments performed on MUs from the medial gastrocnemius muscle of five rats. Besides the easy calculation of the twitch parameters and their variability, the usefulness of the model was confirmed by eliminating artifacts and noise in the original twitch records, as well as by calculations of the velocity of force increase and decrease, the area under force records, and by normalization of all twitches with respect to the maximal force and contraction time. It was concluded that: (1) the four-parameter twitch model describes precisely the individual contractions of various MUs; (2) all physiological twitch parameters are distributed continuously and located within overlapping intervals for different MU types; this distribution is not linear, but exponential; (3) S MUs can be distinguished from fast ones on the basis of some twitch parameters (contraction and half-relaxation times, velocity of contraction), but the same cannot be applied for FF and FR MUs; (4) the analysis of the normalized twitches reveals the differences in shapes for different types of MUs, which shows that twitches of different MUs cannot be obtained from one standard pattern scaled in time and force. These results may have functional implications for studying effectiveness of twitch summation during tetanic contractions and the work performed by various types of MUs.

Modeling of summation of individual twitches into unfused tetanus for various types of rat motor units.

Repeated stimulation of motor units (MUs) causes an increase of the force output that cannot be explained by linear summation of equal twitches evoked by the same stimulation pattern. To explain this phenomenon, an algorithm for reconstructing the individual twitches, that summate into an unfused tetanus is described in the paper. The algorithm is based on an analytical function for the twitch course modeling. The input parameters of this twitch model are lead time, contraction and half-relaxation times and maximal force. The measured individual twitches and unfused tetani at 10, 20, 30 and 40 Hz stimulation frequency of three rat motor units (slow, fast resistant to fatigue and fast fatigable) are processed. It is concluded that: (1) the analytical function describes precisely the course of individual twitches; (2) the summation of equal twitches does not follow the results from the experimentally measured unfused tetani, the differences depend on the type of the MU and are bigger for higher values of stimulation frequency and fusion index; (3) the reconstruction of individual twitches from experimental tetanic records can be successful if the tetanus is feebly fused (fusion index up to 0.7); (4) both the maximal forces and time parameters of individual twitches subtracted from unfused tetani change and influence the course of each tetanus. A discrepancy with respect to the relaxation phase was observed between experimental results and model prediction for tetani with fusion index exceeding 0.7. This phase was predicted longer than the experimental one for better fused tetani. Therefore, a separate series of physiological experiments and then, more complex model are necessary for explanation of this distinction.

Experimental and modelling investigation of learning a fast elbow flexion in the horizontal plane.

Changes in the kinematic and electromyographic characteristics that occur while learning to move as fast as possible have been studied experimentally. Experimental investigation of what happens to the individual motor units (MUs) is more difficult. Access to each MU is impossible, and the recruitment and force developing properties of all individual MUs cannot be known. Thus, what is currently known about MU firing is based on experiments that have recorded relatively few MUs compared to what exists in the entire muscle. A recently developed muscle model (Raikova and Aladjov, 2002, J. Biomechanics, 35, 1123-1135) composed of MUs with different properties can be used for such investigation. The process of learning fast elbow flexion in the horizontal plane was simulated and the results were compared with experimentally measured data. Comparing the simulation results of the very first trial of a particular subject with those of the last trail (at the end of the learning process), it can be concluded that the speed of limb motion and muscle forces increase initially as a result of the more synchronous MUs activation and the increase of firing rate of active MUs. Further improvement necessitated an appreciable reduction in the motor task requirements (i.e. less muscle force and less MUs' activity) set in the computational algorithm by optimization criteria. This forced the next process-inclusion of additional MUs.

The influence of the way the muscle force is modeled on the predicted results obtained by solving indeterminate problems for a fast elbow flexion.

A critical point in models of the human limbs when the aim is to investigate the motor control is the muscle model. More often the mechanical output of a muscle is considered as one musculotendon force that is a design variable in optimization tasks solved predominantly by static optimization. For dynamic conditions, the relationship between the developed force, the length and the contraction velocity of a muscle becomes important and rheological muscle models can be incorporated in the optimization tasks. Here the muscle activation can be a design variable as well. Recently a new muscle model was proposed. A muscle is considered as a mixture of motor units (MUs) with different peculiarities and the muscle force is calculated as a sum of the MUs twitches. The aim of the paper is to compare these three ways for presenting the muscle force. Fast elbow flexion is investigated using a planar model with five muscles. It is concluded that the rheological models are suitable for calculation of the current maximal muscle forces that can be used as weight factors in the objective functions. The model based on MUs has many advantages for precise investigations of motor control. Such muscle presentation can explain the muscle co-contraction and the role of the fast and the slow MUs. The relationship between the MUs activation and the mechanical output is more clear and closer to the reality.

Hierarchical genetic algorithm versus static optimization-investigation of elbow flexion and extension movements.

The applicability of static optimization (and, respectively, frequently used objective functions) for prediction of individual muscle forces for dynamic conditions has often been discussed. Some of the problems are whether time-independent objective functions are suitable, and how to incorporate muscle physiology in models. The present paper deals with a twofold task: (1) implementation of hierarchical genetic algorithm (HGA) based on the properties of the motor units (MUs) twitches, and using multi-objective, time-dependent optimization functions; and (2) comparison of the results of the HGA application with those obtained through static optimization with a criterion "minimum of a weighted sum of the muscle forces raised to the power of n". HGA and its software implementation are presented. The moments of neural stimulation of all MUs are design variables coding the problem in the terms of HGA. The main idea is in using genetic operations to find these moments, so that the sum of MUs twitches satisfies the imposed goals (required joint moments, minimal sum of muscle forces, etc.). Elbow flexion and extension movements with different velocities are considered as proper illustration. It is supposed that they are performed by two extensor muscles and three flexor muscles. The results show that HGA is a suitable means for precise investigation of motor control. Many experimentally observed phenomena (such as antagonistic co-contraction, three-phasic behavior of the muscles during fast movements) can find their explanation by the properties of the MUs twitches. Static optimization is also able to predict three-phasic behavior and could be used as practicable and computationally inexpensive method for total estimation of the muscle forces.