In the afternoon of Tuesday 17th September, three hands-on tutorials will take place in parallel from 14:30 to 17:30.
Tutorial 1: Towards an integrated predictive software map: Practical examples of use of predictive microbiology software tools for food safety and quality
Tutorial 2: Advanced methods in predictive microbiology
Tutorial 3: Topics in quantitative microbial risk assessment using R
Attendance to any tutorial is free-of-charge for those fully registered in the ICPMF11. However, since places are limited, registered participants must indicate in advance which Tutorial they will attend by informing the ICPMF11 Secretariat (icpmf11@ipb.pt).
For those who wish to attend a pre-conference event without taking part in the ICPMF11, fees apply although priority of places will be given to registered participants.
Tutorial 1: Towards an integrated predictive software map: Practical examples of use of predictive microbiology software tools for food safety and quality
Organisers: Fernando Pérez-Rodríguez1, Pablo Fernández2, Alberto Garre2 and Mariem Ellouze3
Chairs: Fernando Pérez-Rodríguez1 and João Barreira4
1University of Cordoba, Spain
2Technical University of Cartagena, Spain
3Nestle, Switzerland
4Polytechnic Institute of Bragança, Portugal
Justification: Predictive Microbiology has reached its maturity, with international recognition and a variety of applications in several scientific areas, food regulation and food industry (i.e., risk assessment, shelf-life determination, HACCP, etc.). One of the largest achievements attained by Predictive Microbiology has been to successfully combine food microbiology insights with mathematical modelling and, more recently, software engineering. Predictive software tools are now a reality, which open up the predictive modelling world to final users who are not familiar with Predictive Microbiology (such as food inspection services, food authorities and food industry), conveying complex concepts in an interpretable and applicative fashion.
The main objective of Tutorial 1 is to present existing software tools from a practical point of view, integrating and combining the main capabilities and strengths from each one. The hands-on session will be organised to address the main challenges and applications in the use of predictive microbiology and risk assessment, covering the whole predictive pipeline from model construction to application in the real world.
Audience:
1) Predictive microbiology practitioners and beginners from industry, academia and government wishing to learn predictive microbiology basics and know how predictive software can be deployed to tackle the main food safety and quality challenges.
2) Food modellers interested in knowing the current software outlook and deepening into the current initiatives taken to integrate predictive software tools.
Schedule:
14:30 – 15:00 Registration and welcome to participants by the Chairs
Introduction to an integrated predictive software map (Mariem Ellouze, Nestle, Switzerland)
15:00 – 15:30 Flash presentations of the software tools by software developers (~5 min each one)*
15:30 – 15:45 Coffee break
15:45 – 16:15 Hands-on session for model construction
16:15 – 16:45 Hands-on session for model integration in software tools
16:45 – 17:20 Hands-on session for model applications
17:20 – 17:30 Feedback from the audience and final remarks
*The hands-on session follow the typical steps in the predictive model pipeline. In each part, a maximum of three software tools will develop, in parallel, examples of application for the corresponding step. The attendees should register in each step and software tool prior to the hands-on session once the flash presentation has been performed.
The software tools to be included in the hands-on session are:
- Bioinactivation: Alberto Garre, PhD (Technical University of Cartagena, Spain)
- Gropin: Prof. Panos Sakandamis, PhD (University of Athens, Greece)
- Combase: Prof. Mark Tamplin, PhD (University of Tasmania, Australia), Tbc
- Sym’previus: Yvan Le Marc, PhD. (Adria, France)
- Microhibro: Prof. Fernando Pérez-Rodríguez, PhD (University of Córdoba, Spain)
Prerequisites:
Attendees should bring their own laptops to enable the development of the hands-on activities.
Tutorial 2: Advanced methods in predictive microbiology
Organisers: Lihan Huang1, Cheng-An (Andy) Hwang1 and Vasco Cadavez2
Chairs: Lihan Huang1 and Ursula Gonzales-Barron2
1 Residue Chemistry and Predictive Microbiology Research Unit, Eastern Regional Research Center, Agricultural Research Service, U.S. Department of Agriculture, USA
2 School of Agriculture, Polytechnic Institute of Bragança, Portugal
Justification: In recent years, significant progress has been made in predictive modelling research and application. Many predictive models, tools, and databases have been developed for data analysis, model development, and risk assessment, and are available to the industry, academia, international organisations, and governments around the world. Many of these technical resources and application tools, available online or on desktop, can provide a fast and reliable decision-making process for food safety and quality in the food industry. Typical applications of predictive microbiology may include prediction of microbial behaviour during food processing and storage, shelf-life prediction, performance and validation of sampling plans, and quantitative risk assessments.
Tutorial 2 will summarise, present, and discuss the most recent developments; demonstrate both the fundamental and applied aspects of predictive microbiology; and introduce the up-to-date one-step dynamic analysis in predictive modelling. This workshop will use the USDA Integrated Pathogen Modeling Program (USDA IPMP) to demonstrate the basic concepts of predictive microbiology and introduce one-step kinetic analysis using the USDA IPMP-Global Fit. More advanced topics, such as dynamic modelling, will be demonstrated using R, an open-source statistical package. Methods in experimental design and sensitivity analysis will also be covered. The aim of this workshop is to discuss practical use of these computing tools to develop accurate predictive models more effectively and efficiently.
Audience:
Students, scientists and engineers who are interested in developing predictive models for microbial shelf-life prediction and risk assessment.
Schedule:
14:30 – 14:45 Welcome to participants and introduction of tutorial and objectives by the Chairs
14:45 – 15:15 Introduction to predictive microbiology, model development, and applications (Andy Hwang)
15:15 – 15:45 Methods in predictive modelling: USDA IPMP and IPMP-Global Fit for research and development (Lihan Huang)
15:45 – 16:00 Coffee break
16:00 – 16:45 Predictive microbiology made easy with R programming (Vasco Cadavez)
16:45 – 17:30 Advanced topics in predictive modelling – experimental design, sensitivity analysis, dynamic modelling, and current research focuses at the USDA: research needs and collaboration (Lihan Huang)
Prerequisites:
For basic concepts: None.
For advanced topics: Some knowledge of R or MATLAB programming is desirable. Attendees should bring their own laptops.
Tutorial 3: Topics in quantitative microbial risk assessment using R
Organisers: Patrick Njage1 and Ana Sofia Ribeiro Duarte1
Chairs: Patrick Njage1 and Elsa Ramalhosa2
1Research Group for Genomic Epidemiology, National Food Institute, Technical University of Denmark
2 School of Agriculture, Polytechnic Institute of Bragança, Portugal
Justification
The introduction of R as an entirely open-sourced data analytics tool has led to the development of which currently is the most comprehensive statistical analysis package available. R continues to incorporate all of the standard statistical tests, models and analyses and a team of developers help in tweaking this software such that most of the latest technological developments, are first to arrive on this software before they are seen anywhere else. However, despite these potential advantages and much historical use of R in epidemiology, the incorporation R in quantitative microbiological risk assessment (QMRA) remains unexplored. Various commercial or open source tools of low flexibility in user manipulation are commonly used. Certain aspects during modelling may affect the validity of QMRA results, while tools to address these aspects have already been developed in R.
This workshop demonstrates the use of R in QMRA. The advantages and flexibility of R will also be demonstrated for selected commonly-overlooked perspectives in data analysis during QMRA. A final part of the workshop will involve a demonstration on machine learning methods as a tool for hazard characterization and source attribution, applying next generation sequencing data.
Audience:
Students, scientists and engineers who are interested in developing flexible quantitative risk assessment models using the R software.
Schedule:
14:30 – 16:00 Uncertainty and variability: Simulation, censored count data and non-normal distributions for microbial concentrations (Poisson and alternative mixture distributions)
16:00 – 16:15 Coffee break
16:15 – 17:10 Quantifying consumption data: correlated inputs, copula theory; dose-response assessment using mechanistic and empirical models; power and advantages of model averaging
17:10 – 17:30 Demonstration on machine learning tools for hazard characterisation and source attribution from next generation sequencing data.
Prerequisites:
Basic ability to use R is desirable. R version 3.2.0 or later ( https://cran.r-project.org/) and Rstudio (https://www.rstudio.com/products/rstudio/download/) must be installed. R codes and other materials for the workshop will be provided for all registered participants.