PERISCOPE-Opt: Machine learning-based prediction of optimal fermentation conditions and yields of recombinant periplasmic protein expressed in Escherichia coli

Kulandai Arockia Rajesh Packiam; Chien Wei Ooi; Fuyi Li; Shutao Mei; Beng Ti Tey; Huey Fang Ong; Jiangning Song; Ramakrishnan Nagasundara Ramanan

doi:10.1016/j.csbj.2022.06.006

PERISCOPE-Opt: Machine learning-based prediction of optimal fermentation conditions and yields of recombinant periplasmic protein expressed in Escherichia coli

Comput Struct Biotechnol J. 2022 Jun 3:20:2909-2920. doi: 10.1016/j.csbj.2022.06.006. eCollection 2022.

Authors

Kulandai Arockia Rajesh Packiam¹, Chien Wei Ooi^{1

2}, Fuyi Li³, Shutao Mei⁴, Beng Ti Tey^{1

2}, Huey Fang Ong⁵, Jiangning Song^{4

6}, Ramakrishnan Nagasundara Ramanan¹

Affiliations

¹ Chemical Engineering Discipline, School of Engineering, Monash University Malaysia, Jalan Lagoon Selatan, 47500 Bandar Sunway, Malaysia.
² Advanced Engineering Platform, Monash University Malaysia, Jalan Lagoon Selatan, 47500 Bandar Sunway, Selangor, Malaysia.
³ Department of Microbiology and Immunology, The Peter Doherty Institute for Infection and Immunity, The University of Melbourne, Victoria 3010, Australia.
⁴ Biomedicine Discovery Institute and Department of Biochemistry and Molecular Biology, Monash University, Victoria 3800, Australia.
⁵ School of Information Technology, Monash University Malaysia, Jalan Lagoon Selatan, 47500 Bandar Sunway, Malaysia.
⁶ Monash Centre for Data Science, Faculty of Information Technoology, Monash University, Victoria 3800, Australia.

Abstract

Optimization of the fermentation process for recombinant protein production (RPP) is often resource-intensive. Machine learning (ML) approaches are helpful in minimizing the experimentations and find vast applications in RPP. However, these ML-based tools primarily focus on features with respect to amino-acid-sequence, ruling out the influence of fermentation process conditions. The present study combines the features derived from fermentation process conditions with that from amino acid-sequence to construct an ML-based model that predicts the maximal protein yields and the corresponding fermentation conditions for the expression of target recombinant protein in the Escherichia coli periplasm. Two sets of XGBoost classifiers were employed in the first stage to classify the expression levels of the target protein as high (>50 mg/L), medium (between 0.5 and 50 mg/L), or low (<0.5 mg/L). The second-stage framework consisted of three regression models involving support vector machines and random forest to predict the expression yields corresponding to each expression-level-class. Independent tests showed that the predictor achieved an overall average accuracy of 75% and a Pearson coefficient correlation of 0.91 for the correctly classified instances. Therefore, our model offers a reliable substitution of numerous trial-and-error experiments to identify the optimal fermentation conditions and yield for RPP. It is also implemented as an open-access webserver, PERISCOPE-Opt (http://periscope-opt.erc.monash.edu).

Keywords: AUC, area under the curve; CV, cross-validation; CfsSubsetEval, Correlation-based Forward Selection Subset Evaluator; ClassifierSubsetEval, Classifier Subset Evaluator; E. coli, Escherichia coli; Escherichia coli; FC1, Feature Category 1; FC2, Feature Category 2; FC3, Feature Category 3; FC4, Feature Category 4; IPTG, isopropyl β-D-1-thiogalactopyranoside; LOOCV, Leave-one-out cross-validation; MAE, mean absolute error; MCC, Mathew correlation coefficient; ML, machine learning; MLR, machine learning in R; Machine learning; OD, optical density at 600 nm; Optimization; PCC, Pearson correlation coefficient; Periplasmic expression; Prediction model; RF, random forest; RFR, RF regression; RFR-High, RFR for high; RFR-Medium, RFR for medium; RMSE, root mean squared error; RPP, Recombinant protein production; RSM, response surface methodology; Recombinant protein production; SMOTE, Synthetic Minority Over-sampling Technique; SP, signal peptides; SVM, support vector machines; SVR, SVM regression; SVR-Low, SVR for class: "low"; XGB, XGBoost; pI, isoelectric point.