Novel calibration design improves knowledge transfer across products for the characterization of pharmaceutical bioprocesses.

Modern machine learning has the potential to fundamentally change the way bioprocesses are developed. In particular, horizontal knowledge transfer methods, which seek to exploit data from historical processes to facilitate process development for a new product, provide an opportunity to rethink current workflows. In this work, we first assess the potential of two knowledge transfer approaches, meta learning and one-hot encoding, in combination with Gaussian process (GP) models. We compare their performance with GPs trained only on data of the new process, that is, local models. Using simulated mammalian cell culture data, we observe that both knowledge transfer approaches exhibit test set errors that are approximately halved compared to those of the local models when two, four, or eight experiments of the new product are used for training. Subsequently, we address the question whether experiments for a new product could be designed more effectively by exploiting existing knowledge. In particular, we suggest to specifically design a few runs for the novel product to calibrate knowledge transfer models, a task that we coin calibration design. We propose a customized objective function to identify a set of calibration design runs, which exploits differences in the process evolution of historical products. In two simulated case studies, we observed that training with calibration designs yields similar test set errors compared to common design of experiments approaches. However, the former requires approximately four times fewer experiments. Overall, the results suggest that process development could be significantly streamlined when systematically carrying knowledge from one product to the next.

An innovative hybrid modeling approach for simultaneous prediction of cell culture process dynamics and product quality.

The use of hybrid models is extensively described in the literature to predict the process evolution in cell cultures. These models combine mechanistic and machine learning methods, allowing the prediction of complex process behavior, in the presence of many process variables, without the need to collect a large amount of data. Hybrid models cannot be directly used to predict final product critical quality attributes, or CQAs, because they are usually measured only at the end of the process, and more mechanistic knowledge is needed for many classes of CQAs. The historical models can instead predict the CQAs better; however, they cannot directly relate manipulated process parameters to final CQAs, as they require knowledge of the process evolution. In this work, we propose an innovative modeling approach based on combining a hybrid propagation model with a historical data-driven model, that is, the combined hybrid model, for simultaneous prediction of full process dynamics and CQAs. The performance of the combined hybrid model was evaluated on an industrial dataset and compared to classical black-box models, which directly relate manipulated process parameters to CQAs. The proposed combined hybrid model outperforms the black-box model by 33% on average in predicting the CQAs while requiring only around half of the data for model training to match performance. Thus, in terms of model accuracy and experimental costs, the combined hybrid model in this study provides a promising platform for process optimization applications.

No evidence of attentional prioritization for threatening targets in visual search.

Throughout human evolutionary history, snakes have been associated with danger and threat. Research has shown that snakes are prioritized by our attentional system, despite many of us rarely encountering them in our daily lives. We conducted two high-powered, pre-registered experiments (total N = 224) manipulating target prevalence to understand this heightened prioritization of threatening targets. Target prevalence refers to the proportion of trials wherein a target is presented; reductions in prevalence consistently reduce the likelihood that targets will be found. We reasoned that snake targets in visual search should experience weaker effects of low target prevalence compared to non-threatening targets (rabbits) because they should be prioritized by searchers despite appearing rarely. In both experiments, we found evidence of classic prevalence effects but (contrasting prior work) we also found that search for threatening targets was slower and less accurate than for nonthreatening targets. This surprising result is possibly due to methodological issues common in prior studies, including comparatively smaller sample sizes, fewer trials, and a tendency to exclusively examine conditions of relatively high prevalence. Our findings call into question accounts of threat prioritization and suggest that prior attention findings may be constrained to a narrow range of circumstances.

Unveiling the Borohydride Ion through Force-Field Development.

The borohydride ion, BH4-, is an essential reducing agent in many technological processes, yet its full understanding has been elusive, because of at least two significant challenges. One challenge arises from its marginal stability in aqueous solutions outside of basic pH conditions, which considerably limits the experimental thermodynamic data. The other challenge comes from its unique and atypical hydration shell, stemming from the negative excess charge on its hydrogen atoms, which complicates the accurate modeling in classical atomistic simulations. In this study, we combine experimental and computer simulation techniques to devise a classical force field for NaBH4 and deepen our understanding of its characteristics. We report the first measurement of the ion's activity coefficient and extrapolate it to neutral pH conditions. Given the difficulties in directly measuring its solvation free energies, owing to its instability, we resort to quantum chemistry calculations. This combined strategy allows us to derive a set of nonpolarizable force-field parameters for the borohydride ion for classical molecular dynamics simulations. The derived force field simultaneously captures the solvation free energy, the hydration structure, as well as the activity coefficient of NaBH4 salt across a broad concentration range. The obtained insights into the hydration shell of the BH4- ion are crucial for accurately modeling and understanding its interactions with other molecules, ions, materials, and interfaces.

Human emotional evaluation of ancestral and modern threats: fear, disgust, and anger.

Introduction: Animal and human ancestors developed complex physiological and behavioral response systems to cope with two types of threats: immediate physical harm from predators or conspecifics, triggering fear, and the risk of infections from parasites and pathogens leading to the evolution of the behavioral immune system with disgust as the key emotion. Integration of the evolutionary concepts of the fear module and behavioral immune systems has been infrequent, despite the significant survival advantages of disgust in various contexts. Studies comparing attention to ancestral and modern threats accompanied by fear have yielded ambiguous results and what qualifies as salient modern disgusting stimuli remains unclear. We do not know whether disgust or the behavioral immune system, as inherent aspects of human psychology, have adapted to safeguard us from pandemic risks or poisoning by modern toxic substances. Methods: To test these effects, we have developed a survey comprised of 60 short vignettes describing threats evoking fear and disgust belonging to one of the three main categories of threats: (1) ancestral (phylogenetic), (2) modern (ontogenetic), and (3) pandemics of airborne disease. Each vignette was evaluated on a 7-point Likert scale based on fear, disgust, and anger. In total, 660 respondents completed the survey. The data were analysed using a factor analysis and general linear model with the respondent as a random factor. Results: The results show that the strongest fear is triggered by modern threats (electricity, car accidents), while the highest disgust is evoked by ancient threats (body waste products, worms, etc.). Interestingly, disgust does not respond to modern threat stimuli such as toxic substances or radioactivity as these evoke mainly fear and anger. Finally, a distinct response pattern was found for pandemic threats, in which both fear (e.g., of disease and death) and disgust (e.g., of used face masks) are employed. Discussion: Our study offers valuable insights into the emotional responses to ancestral and modern threats and their adaptation to pandemic challenges. Ancestral threats are not always more powerful stimuli than adequate threats of the modern type, but they function specifically. Thus, snakes and heights as fear-inducing ancestral threats form separate factors in a multivariate analysis, whereas all ancestral disgust stimuli group together. The threat of a pandemic forms a specific category and people process it emotionally and cognitively. These insights contribute to our understanding of human psychology and behavior in an ever-changing world.