Search

11th International Conference on Practical Applications of Computational Biology & Bioinformatics held in Porto, Portugal, 21—23 June 2017, In this work we describe a new methodology to improve predictive capabilities of dynamic models when parameters differ in orders of magnitude. The main idea is to normalise the model unknown parameters before solving the classical problem of optimal experimental design based on the Fisher information matrix. The normalisation improves the relative confidence intervals of the estimated parameters and the conditioning of the Fisher matrix, especially for those criteria aiming to decorrelate the model parameters. Using the so-called core predictions, we show how the new approach improves the final model predictive capabilities in two terms: predictions are closer to the real dynamics and with better confidence intervals. We illustrate the concepts using two toy examples linear and non-linear in their parameters. Finally we test the performance of the normalisation in a model simulating the bacterial SOS response. This pathway remains of main relevance to work towards a predictive model of antimicrobial resistance

11 páginas, 4 tablas, 9 figuras, Fish wastage and market prices highly depend on accurate and reliable predictions of product shelf life and quality. The Quality Index Method (QIM) and EU grading criteria for whitefish (Council Regulation(EC) No 2406/96, 1996) are established sensory methods used in the market to monitor fish quality. Each assessment requires the consultation of a panel of trained experts. The indexes refer exclusively to the current state of the fish without any predictions about its evolution in the following days. This work proposes the development of a smart quality sensor which enables to measure quality and to predict its progress through time. The sensor combines information of biochemical and microbial spoilage indexes with dynamic models to predict quality in terms of the QIM and EU grading criteria. Besides, the sensor can account for the variability inside the batch if spoilage indexes are measured in more than one fish sample. The sensor is designed and tested to measure quality in fresh cod (Gadus morhua) under commercial ice storage conditions. Only two spoilage indexes, psychrotrophic counts and total volatile base-nitrogen content, were required to get accurate estimations of the two usual established sensory methods. The sensor is able to account for biological variability as shown with the validation and demonstration data sets. Moreover, new research and technologies are in course to make these measurements faster and non-destructive. This would allow having at hand a smart non-intrusive fish quality sensor, This work has been funded by the Spanish Ministry of Science and Innovation throughout project ISFORQUALITY (AGL2012- 39951-C02-01), RESISTANCE (DPI2014-54085-JIN) and FP7 SPEC- TRAFISH project (FP7/2007-2013: grant no 605399)

One of the most important properties of a mathematical model is the ability to make predictions: to predict that which has not yet been measured. Such predictions can sometimes be obtained from a simple simulation, but that requires that the parameters in the model are known from before. In biology, the parameters are usually both not known from before and not identifiable, i.e. the parameter values cannot be determined uniquely from available data. In such cases of unidentifiability, the space of acceptable parameters is large, often infinite in certain directions. For such large spaces, sampling-based approaches that try to characterize the entire space have difficulties. Recently, a new type of alternative approaches that circumvent this characterization problem has been proposed: where one only searches those directions in the space of acceptable parameters that are relevant for the uncertainty of a particular prediction. In this review chapter, these recently proposed methods are compared and contrasted, both regarding theoretical properties, and regarding user experience. The focus is on methods from the field of systems biology, but also methods from biostatistics, pharmacodynamics, and biochemometrics are discussed. The hope is that this review will increase the usefulness and understanding of already proposed methods, and thereby help foster a tradition where predictions only are deemed interesting if their uncertainties have been determined.

1 póster.-- 29th EFFoST International Conference, 10-12 November 2015, Athens, Greece, Reduction of fish freshness during the supply chain results into severe losses both from the economic (drastic reductions on market price) and environmental (fish rejection) points of view. In this context, prediction techniques based on mathematical models emerge as useful tools for fish quality assessment. In this work we propose a methodology to forecast fish freshness during storage. In this regard, we found a strong correlation between fish freshness indicators commonly used in the industry, such as the Quality Sensory Method (QSM), and the growth of Specific Spoilage Organisms (SSO). The main idea is to develop a reliable mathematical model which, on the one hand, describes the growth of SSO during storage and, on the other hand, relates the QSM to SSO concentration. Issues such as storage conditions, fishing practices or gutting procedures are taken into account. Such model allows us to predict shelf-life and quality losses. In order to provide the model with reliable predictive capabilities, advanced model calibration techniques are employed. In this regard, optimal experimental design (OED) is used to generate sufficient nonisothermal informative experiments that allow us to reduce the uncertainty in the predictions. Also, the Monte Carlo sampling technique is exploited to detect the parameter confidence intervals and to calculate fish-to-fish variability in the predictions (core predictions). The approach is illustrated using hake (Merluccius merluccius) caught either by bottom-set net or long-line under different storage temperature conditions. Furthermore, the effect of gutting is also studied. This methodology is flexible enough to consider other fish species or stress variables such as atmosphere composition, This work has been funded by the LIFE EU program: project LIFE iSEAS (LIFE13 ENV/ES/000131); by the Spanish Ministry of Science and Innovation: projects ISFORQUALITY (AGL2012-39951-C02-01), RESISTANCE (DPI2014-54085-JIN); and by Xunta de Galicia: contrato programa 2015

10 páginas, 5 tablas, 6 figuras.-- This is an open access article under the CC BY-NC-ND license, Fish quality has a direct impact on market price and its accurate assessment and prediction are of main importance to set prices, increase competitiveness, resolve conflicts of interest and prevent food wastage due to conservative product shelf-life estimations. In this work we present a general methodology to derive predictive models of fish freshness under different storage conditions. The approach makes use of the theory of optimal experimental design, to maximize data information and in this way reduce the number of experiments. The resulting growth model for specific spoilage microorganisms in hake (Merluccius merluccius) is sufficiently informative to estimate quality sensory indexes under time-varying temperature profiles. In addition it incorporates quantitative information of the uncertainty induced by fish variability. The model has been employed to test the effect of factors such as fishing gear or evisceration, on fish spoilage and therefore fish quality. Results show no significant differences in terms of microbial growth between hake fished by long-line or bottom-set nets, within the implicit uncertainty of the model. Similar conclusions can be drawn for gutted and un-gutted hake along the experiment horizon. In addition, whenever there is the possibility to carry out the necessary experiments, this approach is sufficiently general to be used in other fish species and under different stress variables, This work has been funded by the Spanish Ministry of Science and Innovation throughout project ISFORQUALITY (AGL2012-39951-C02-01) and by the CSIC throughout project CONTROLA (PIE201270E075)

Fish quality has a direct impact on market price and its accurate assessment and prediction is of main importance to prevent food wastage due to extremely conservative product shelf-life estimations. In this work we present a general methodology to derive predictive models of fish freshness under different storage conditions. The approach focuses in finding growth models of specific spoilage micro-organisms in hake sufficiently informative to estimate two common quality market indexes. In order to get enough predictive capabilities, we make use of Optimal Experimental Design and analyze the statistics associated with the predictions due to fish variability. As a result, both quality indexes and their variability can be estimated under different temperature profiles. Although the methodology here presented has been applied and validated in hake, it is sufficiently general to be used in other fish species and under different stress variables. In addition, the parameters found for this case are robust enough to predict quality in conditions not used during the model calibration, such as for example hake captured with a different commercial fishing technique.