Publications

What is a Publication?
4 Publications visible to you, out of a total of 4
MIFE and MIFD: Minimum information for fermentation experiments and devices

Abstract (Expand)

Abstract Background As the technological advancements of the early 21st century are pushing industrial biotechnology (IB) into the realm of Big Data–driven innovation, the requirement for trustworthy …Big Data–driven innovation, the requirement for trustworthy data management, annotation, and standardization is emerging as a necessity. Minimum information models (MIMs) have long been used across disciplines as the backbone of good data management practices by providing the scaffold upon which standardized recording of metadata can adequately and succinctly describe an understudied phenomenon. Findings Here we present a minimum set of metadata, named the minimum information for fermentation experiments (MIFE) and devices (MIFD), that has been specifically designed to accommodate the data management and annotation needs of IB-related fermentation experiments. Although the proposed schema is tailored to IB applications, MIFE and MIFD build upon well-established models and community standards to facilitate easier integration with existing infrastructure and easier adoption by the community, as well as aim to integrate Findable, Accessible, Interoperable, and Reproducible (FAIR) principles in the IB field. In addition, the integration with FAIR Data Station (FAIR DS), a tool that offers metadata validation and enables the automated uptake of (meta)data from data management repositories such as FAIRDOM-SEEK, is showcased. The proposed models are accompanied by a Python package that enables their programmatic use by creating a Linked Data Modeling Language (LinkML) schema that can fuel subsequent analyses. Conclusions Through the promotion and simplification of knowledge discovery, we believe that MIFE and MIFD can accelerate the application of state-of-the-art artificial intelligence (AI) methods and the adoption of explainable AI to better understand bioprocesses at scale.

Authors: Georgios K Georgakilas, Brett Metcalfe, Ariane Bize, Matthew Crowther, Emilie Fernandez, Susana Maria Alonso Villela, Stuart Owen, Rudolf Wittner, David Camilo Corrales, Anselm von Gladiss, Peter Blomberg, Munazah Andrabi, Cesar Arturo Aceves Lara, Hans Mattila, Marily Wiebe, Theodore Dalamagas, Jasper J Koehorst

Date Published: 2026

Publication Type: Journal

Soft sensors based on interpretable learners for industrial-scale fed-batch fermentation: Learning from simulations

Abstract (Expand)

Monitoring bioprocesses is a challenging task where most of the variables of interest can only be measured offline. Soft sensors have emerged as a solution to provide online estimations. This work … compares interpretable learners such as CART, M5, CUBIST, and Random Forest as soft sensors for industrial-scale fed-batch fermentation of penicillin production. A structured model of industrial-scale penicillin fermentation is implemented to generate the dataset and train the interpretable learners. Variables such as substrate feed rate, agitation, temperature, pH, dissolved oxygen, vessel volume, CO2, and O2 percent in off-gas are considered as independent (predictors). The CUBIST model has achieved the best results with values of 0.908, 9.916, 3.149, and 1.920 for the coefficient of determination, Mean Squared Error, Root Mean Squared Error, and Mean Absolute Error, respectively. These results demonstrate the feasibility of developing soft sensors based on interpretable models to predict penicillin concentration at an industrial scale.

Authors: Juan Camilo Acosta-Pavas, Carlos Eduardo Robles-Rodriguez, David Griol, Fayza Daboussi, Cesar Arturo Aceves-Lara, David Camilo Corrales

Date Published: 1st Aug 2024

Publication Type: Journal

STAMM: Soft sensor moniToring and mAintenance framework for Machine learning Models

Abstract (Expand)

STAMM (Soft sensor moniToring and mAintenance framework for Machine learning Models) is an open-source MLOps framework for industrial machine-learning soft sensors. Unlike general-purpose MLOps platforms … (MLflow, Kubeflow, Metaflow, ClearML), STAMM targets the regime where ground-truth labels arrive offline hours or days late, processes exhibit slow non-stationary dynamics, and models are multi-language. The framework comprises five loosely coupled components: a time-series database, workflow orchestrator, language-agnostic REST model registry, dashboard with human-in-the-loop labelling, and an extensible drift-detection package. STAMM was validated on an industrial-scale fed-batch penicillin fermentation (IndPenSim) with seven coexisting R and Python soft sensors served through the model registry.

Authors: Carlos Suarez, Alexander Astudillo, Brett Metcalfe, Matthew Crowther, Jasper J. Koehorst, Esteban Castillo, Ariane Bize, David Camilo Corrales

Date Published: 1st Sep 2026

Publication Type: Journal

Towards a machine learning operations (MLOps) soft sensor for real-time predictions in industrial-scale fed-batch fermentation

Abstract (Expand)

Real-time predictions in fermentation processes are crucial because they enable continuous monitoring and control of bioprocessing. However, the availability of online measurements is limited by the … availability and feasibility of sensing technology. Soft sensors - or software sensors that convert available measurements into measurements of interest (product yield, quality, etc.) - have the potential to improve efficiency and product quality. Machine learning (ML) based soft sensors have gained increased popularity over the years since they can incorporate variables that are measured in real-time, and exploit the intricate patterns embedded in such voluminous datasets. However, ML-based soft sensor requires more than just a classical ML learner with an unseen test set to evaluate the quality prediction of the model. When a ML model is deployed in production, its performance can deteriorate rapidly leading to an unanticipated decline in the quality of the output and predictions. Here a proof concept of Machine Learning Operations (MLOps) to automate the end-to-end soft sensor lifecycle in industrial scale fed-batch fermentation, from development and deployment to maintenance and monitoring is proposed. Using the industrial-scale penicillin fermentation (IndPenSim) dataset that includes 100 fermentation batches, to build a soft sensor based on Long Short Term Memory (LSTM) for penicillin concentration prediction. The batches containing deviations in the processes (91–100) were used to assess concept drift of the LSTM soft sensor. The evaluation of concept drift is evidenced by the soft sensor performance falling below the set threshold based on the Population Stability Index (PSI), which automatically triggers an alert to run the retraining pipeline.

Authors: Brett Metcalfe, Juan Camilo Acosta-Pavas, Carlos Eduardo Robles-Rodriguez, George K. Georgakilas, Theodore Dalamagas, Cesar Arturo Aceves-Lara, Fayza Daboussi, Jasper J Koehorst, David Camilo Corrales

Date Published: 1st Mar 2025

Publication Type: Journal

Powered by
 (v.1.17.4)
About IBISBA | About IBISBA Knowledge Hub | Funding and Programmes | Credits | Terms & Conditions
Copyright © 2008 - 2026 The University of Manchester and HITS gGmbH
IBISBA is a pan-European research infrastructure currently funded through multiple EU projects: IBISBA 1.0 (H2020 grant agreement No. 730976), PREP-IBISBA (H2020 grant agreement No. 871118) and the follow-on project IBISBA-DIALS (grant agreement No. 101131085) and BIOINDUSTRY 4.0 (Grant agreement No. 101094287). Registering data or other knowledge assets on this platform is the sole responsibility of Users. IBISBA cannot be held responsible for misuse or misappropriation of data and assets belonging to a Third Party.