An artificial neural network classification method employing longitudinally monitored immune biomarkers to predict the clinical outcome of critically ill COVID-19 patients

Abstract

Background The severe form of COVID-19 can cause a dysregulated host immune syndrome that might lead patients to death. To understand the underlying immune mechanisms that contribute to COVID-19 disease we have examined 28 different biomarkers in two cohorts of COVID-19 patients, aiming to systematically capture, quantify, and algorithmize how immune signals might be associated to the clinical outcome of COVID-19 patients. Methods The longitudinal concentration of 28 biomarkers of 95 COVID-19 patients was measured. We performed a dimensionality reduction analysis to determine meaningful biomarkers for explaining the data variability. The biomarkers were used as input of artificial neural network, random forest, classification and regression trees, k-nearest neighbors and support vector machines. Two different clinical cohorts were used to grant validity to the findings. Results We benchmarked the classification capacity of two COVID-19 clinicals studies with different models and found that artificial neural networks was the best classifier. From it, we could employ different sets of biomarkers to predict the clinical outcome of COVID-19 patients. First, all the biomarkers available yielded a satisfactory classification. Next, we assessed the prediction capacity of each protein separated. With a reduced set of biomarkers, our model presented 94% accuracy, 96.6% precision, 91.6% recall, and 95% of specificity upon the testing data. We used the same model to predict 83% and 87% (recovered and deceased) of unseen data, granting validity to the results obtained. Conclusions In this work, using state-of-the-art computational techniques, we systematically identified an optimal set of biomarkers that are related to a prediction capacity of COVID-19 patients. The screening of such biomarkers might assist in understanding the underlying immune response towards inflammatory diseases.

Model-interpreted outcomes of artificial neural networks classifying immune biomarkers associated with severe infections in ICU

“…In fact, the robustness of this method was previously reported ( 24 ) as a solid way to find patterns in tabular data, among others. Additionally, the appropriate selection of the input variables is a key process in obtaining satisfactory results ( 25 ). In many cases, classification and regression models derived from lower-dimensional datasets benefit the downstream decision-making process ( 26 , 27 ).…”

Identification of Marker Genes in Infectious Diseases from ScRNA-seq Data Using Interpretable Machine Learning

“…The analysis of genomics data might be powered by artificial intelligence (AI) techniques, which hold the mathematical power to untangle the intricate relationship between data points in higher-dimensional settings [ 9 ]. Examples of AI-assisted medicine can be found in the identification of marker genes [ 10 , 11 ], biomarkers [ 12 , 13 ], vaccine candidates [ 14 , 15 ], and potential targets for therapeutics [ 16 ]. Moreover, AI has been used in conjunction with scRNA-seq data analysis for the identification of tumor cells [ 17 ].…”

Convergence in Mobility Data Sets From Apple, Google, and Meta

“…Data are commonly seen as abstractions of the real world. For instance, in applications that benefit from data, there is generally a need to promote validation with external data [ 10 ]; this reality is widely explored in clinical settings [ 11 ], genomic studies [ 12 ], among others. Consequently, a model that can react similarly upon different inputs strengthens the conclusions that are derived from such data.…”