Skip to main content
Fig. 1 | Bioresources and Bioprocessing

Fig. 1

From: A novel thermophilic chitinase directly mined from the marine metagenome using the deep learning tool Preoptem

Fig. 1

Construction of a deep learning model to screen for potential thermophilic proteins. A Dataset collected in this study. The x-axis represents species sorted by optimal temperature and y-axis indicates optimal temperature. B Venn diagrams of the psychrophilic (blue), mesophilic (green), and thermophilic (pink) proteins. C Optimization of the parameters of kernel size and number of neurons in the deep learning model. Star indicates the highest value of correlation coefficient. D Correlation between the predicted and observed optimal temperatures in the validation dataset. E Correlation between the predicted and observed optimal temperatures in the test dataset. F Correlation analysis of predicted optimal temperatures for protein with known optimal temperature in the Uniprot database. G Correlation between the predicted and observed optimal temperatures in the cross-validation system. H Correlation between the predicted and observed optimal temperatures for a protein that is not included in the training dataset. I Correlation analysis of predicted optimal temperatures for protein activity and experimentally determined Tm. DI The blue dot indicates that the data are close to the fitted line, the red dot indicates that the data are far away from the fitted line, and the green dot indicates that the data are centered from the fitted line

Back to article page