The ability of blood transcriptome analysis to identify dysregulated pathways and outcome-related genes following myocardial infarction remains unknown. Two gene expression datasets (GSE60993 and GSE61144) were downloaded from Gene Expression Omnibus (GEO) Datasets to identify altered plasma transcriptomes in patients with ST-segment elevated myocardial infarction (STEMI) undergoing primary percutaneous coronary intervention. GEO2R, Gene Ontology/Kyoto Encyclopedia of Genes and Genomes annotations, protein–protein interaction analysis, etc., were adopted to determine functional roles and regulatory networks of differentially expressed genes (DEGs). Dysregulated expressomes were verified at transcriptional and translational levels by analyzing the GSE49925 dataset and our own samples, respectively. A total of 91 DEGs were identified in the discovery phase, consisting of 15 downregulated genes and 76 upregulated genes. Two hub modules consisting of 12 hub genes were identified. In the verification phase, six of the 12 hub genes exhibited the same variation patterns at the transcriptional level in the GSE49925 dataset. Among them, S100A12 was shown to have the best discriminative performance for predicting in-hospital mortality and to be the only independent predictor of death during follow-up. Validation of 223 samples from our center showed that S100A12 protein level in plasma was significantly lower among patients who survived to discharge, but it was not an independent predictor of survival to discharge or recurrent major adverse cardiovascular events after discharge. In conclusion, the dysregulated expression of plasma S100A12 at the transcriptional level is a robust early prognostic factor in patients with STEMI, while the discrimination power of the protein level in plasma needs to be further verified by large-scale, prospective, international, multicenter studies.

No abstract

No abstract

A methodology of data-driven damage state quantification with a probability estimation for structural hysteresis of RC columns is presented in this paper. The knowledge learned from a large-volume structural behavior database is fully considered for developing monitoringoriented damage indicators, with the established relationship between data-driven damage prediction and physics-based damage state evaluation. In the present study, a database of the hysteresis behavior of 1015 RC columns with different design parameters is first generated by adopting OpenSees, with categorization according to the primary design parameter of the axial load ratio. Four limit states of seismic performance with the corresponding values of the proposed damage index are calculated to generate an informative mapping between critical damage states and damage index values. By fitting probabilistic models on the grouped data of damage indices, the exceeding probabilities of damage states corresponding to damage index values can be obtained. Illustrative examples of full-scale RC columns with cyclic loading and shaking table tests are adopted to illustrate the proposed performance-based damage evaluation process.

No abstract