A carregar...
Projeto de investigação
Not Available
Financiador
Autores
Publicações
A general deep hybrid model for bioreactor systems
Publication . Pinto, José; Mestre, Mykaella; Ramos, J.; Costa, Rafael S.; Striedner, Gerald; Oliveira, Rui; LAQV@REQUIMTE; PERGAMON-ELSEVIER SCIENCE LTD
Numerous studies have reported the use of hybrid semiparametric systems that combine shallow neural networks with First Principles for bioprocess modeling. Here we revisit the general bioreactor hybrid model and introduce some deep learning techniques. Multi-layer networks with varying depths were combined with First Principles equations in the form of deep hybrid models. Deep learning techniques, namely the adaptive moment estimation method (ADAM), stochastic regularization and depth-dependent weights initialization were evaluated in a hybrid modeling context. Modified sensitivity equations are proposed for the computation of gradients in order to reduce CPU time for the training of deep hybrid models. The methods are illustrated with applications to a synthetic dataset and a pilot 50 L MUT+ Pichia pastoris process expressing a single chain antibody fragment. All in all, the results point to a systematic generalization improvement of deep hybrid models over its shallow counterpart. Moreover, the CPU cost to train the deep hybrid models is shown to be lower than for the shallow counterpart. In the pilot 50L MUT+ Pichia pastoris data set, the prediction accuracy was increased by 18.4% and the CPU decreased by 43.4%.
Hybrid deep modeling of a CHO-K1 fed-batch process
Publication . Pinto, José; Ramos, João R. C.; Costa, Rafael S.; Rossell, Sergio; Dumas, Patrick; Oliveira, Rui; LAQV@REQUIMTE; DQ - Departamento de Química; Frontiers Media
Hybrid modeling combining First-Principles with machine learning is becoming a pivotal methodology for Biopharma 4.0 enactment. Chinese Hamster Ovary (CHO) cells, being the workhorse for industrial glycoproteins production, have been the object of several hybrid modeling studies. Most previous studies pursued a shallow hybrid modeling approach based on threelayered Feedforward Neural Networks (FFNNs) combined with macroscopic material balance equations. Only recently, the hybrid modeling field is incorporating deep learning into its framework with significant gains in descriptive and predictive power.
SBML2HYB
Publication . Pinto, José; Costa, Rafael S.; Alexandre, Leonardo; Ramos, João; Oliveira, Rui; LAQV@REQUIMTE; DQ - Departamento de Química; Oxford University Press
Here we present sbml2hyb, an easy-to-use standalone Python tool that facilitates the conversion of existing mechanistic models of biological systems in Systems Biology Markup Language (SBML) into hybrid semiparametric models that combine mechanistic functions with machine learning (ML). The so-formed hybrid models can be trained and stored back in databases in SBML format. The tool supports a user-friendly export interface with an internal format validator. Two case studies illustrate the use of the sbml2hyb tool. Additionally, we describe HMOD, a new model format designed to support and facilitate hybrid models building. It aggregates the mechanistic model information with the ML information and follows as close as possible the SBML rules. We expect the sbml2hyb tool and HMOD to greatly facilitate the widespread usage of hybrid modeling techniques for biological systems analysis.
Predictability of COVID-19 hospitalizations, intensive care unit admissions, and respiratory assistance in Portugal
Publication . Patrício, André; Costa, Rafael S.; Henriques, Rui; LAQV@REQUIMTE; JMIR Publications
Background: In the face of the current COVID-19 pandemic, the timely prediction of upcoming medical needs for infected individuals enables better and quicker care provision when necessary and management decisions within health care systems. Objective: This work aims to predict the medical needs (hospitalizations, intensive care unit admissions, and respiratory assistance) and survivability of individuals testing positive for SARS-CoV-2 infection in Portugal. Methods: A retrospective cohort of 38,545 infected individuals during 2020 was used. Predictions of medical needs were performed using state-of-the-art machine learning approaches at various stages of a patient's cycle, namely, at testing (prehospitalization), at posthospitalization, and during postintensive care. A thorough optimization of state-of-the-art predictors was undertaken to assess the ability to anticipate medical needs and infection outcomes using demographic and comorbidity variables, as well as dates associated with symptom onset, testing, and hospitalization. Results: For the target cohort, 75% of hospitalization needs could be identified at the time of testing for SARS-CoV-2 infection. Over 60% of respiratory needs could be identified at the time of hospitalization. Both predictions had >50% precision. Conclusions: The conducted study pinpoints the relevance of the proposed predictive models as good candidates to support medical decisions in the Portuguese population, including both monitoring and in-hospital care decisions. A clinical decision support system is further provided to this end.
Predicting metabolic fluxes from omics data via machine learning
Publication . Gonçalves, Daniel M.; Henriques, Rui; Costa, Rafael S.; LAQV@REQUIMTE; DQ - Departamento de Química; Elsevier BV
The accurate prediction of phenotypes in microorganisms is a main challenge for systems biology. Genome-scale models (GEMs) are a widely used mathematical formalism for predicting metabolic fluxes using constraint-based modeling methods such as flux balance analysis (FBA). However, they require prior knowledge of the metabolic network of an organism and appropriate objective functions, often hampering the prediction of metabolic fluxes under different conditions. Moreover, the integration of omics data to improve the accuracy of phenotype predictions in different physiological states is still in its infancy. Here, we present a novel approach for predicting fluxes under various conditions. We explore the use of supervised machine learning (ML) models using transcriptomics and/or proteomics data and compare their performance against the standard parsimonious FBA (pFBA) approach using case studies of Escherichia coli organism as an example. Our results show that the proposed omics-based ML approach is promising to predict both internal and external metabolic fluxes with smaller prediction errors in comparison to the pFBA approach. The code, data, and detailed results are available at the project's repository [1].
Unidades organizacionais
Descrição
Palavras-chave
Contribuidores
Financiadores
Entidade financiadora
Fundação para a Ciência e a Tecnologia
Programa de financiamento
CEEC IND 2017
Número da atribuição
CEECIND/01399/2017/CP1462/CT0015