Dataset Information

Identifying and characterizing high-risk clusters in a heterogeneous ICU population with deep embedded clustering.

ABSTRACT: Critically ill patients constitute a highly heterogeneous population, with seemingly distinct patients having similar outcomes, and patients with the same admission diagnosis having opposite clinical trajectories. We aimed to develop a machine learning methodology that identifies and provides better characterization of patient clusters at high risk of mortality and kidney injury. We analysed prospectively collected data including co-morbidities, clinical examination, and laboratory parameters from a minimally-selected population of 743 patients admitted to the ICU of a Dutch hospital between 2015 and 2017. We compared four clustering methodologies and trained a classifier to predict and validate cluster membership. The contribution of different variables to the predicted cluster membership was assessed using SHapley Additive exPlanations values. We found that deep embedded clustering yielded better results compared to the traditional clustering algorithms. The best cluster configuration was achieved for 6 clusters. All clusters were clinically recognizable, and differed in in-ICU, 30-day, and 90-day mortality, as well as incidence of acute kidney injury. We identified two high mortality risk clusters with at least 60%, 40%, and 30% increased. ICU, 30-day and 90-day mortality, and a low risk cluster with 25-56% lower mortality risk. This machine learning methodology combining deep embedded clustering and variable importance analysis, which we made publicly available, is a possible solution to challenges previously encountered by clustering analyses in heterogeneous patient populations and may help improve the characterization of risk groups in critical care.

SUBMITTER: Castela Forte J

PROVIDER: S-EPMC8187398 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:BACKGROUND:Emerging evidence has revealed that some long intergenic non-coding RNAs (lincRNAs) are likely to form clusters on the same chromosome, and lincRNA genomic clusters might play critical roles in the pathophysiological mechanism. However, the comprehensive investigation of lincRNA clustering is rarely studied, particularly the characterization of their functional significance across different cancer types. METHODS:In this study, we firstly constructed a computational method basing a sliding window approach for systematically identifying lincRNA genomic clusters. We then dissected these lincRNA genomic clusters to identify common characteristics in cooperative expression, conservation among divergent species, targeted miRNAs, and CNV frequency. Next, we performed comprehensive analyses in differentially-expressed patterns and overall survival outcomes for patients from The Cancer Genome Atlas (TCGA) and The Genotype-Tissue Expression (GTEx) across multiple cancer types. Finally, we explored the underlying mechanisms of lincRNA genomic clusters by functional enrichment analysis, pathway analysis, and drug-target interaction. RESULTS:We identified lincRNA genomic clusters according to the algorithm. Clustering lincRNAs tended to be co-expressed, highly conserved, targeted by more miRNAs, and with similar deletion and duplication frequency, suggesting that lincRNA genomic clusters may exert their effects by acting in combination. We further systematically explored conserved and cancer-specific lincRNA genomic clusters, indicating they were involved in some important mechanisms of disease occurrence through diverse approaches. Furthermore, lincRNA genomic clusters can serve as biomarkers with potential clinical significance and involve in specific pathological processes in the development of cancer. Moreover, a lincRNA genomic cluster named Cluster127 in DLK1-DIO3 imprinted locus was discovered, which contained MEG3, MEG8, MEG9, MIR381HG, LINC02285, AL132709.5, and AL132709.1. Further analysis indicated that Cluster127 may have the potential for predicting prognosis in cancer and could play their roles by participating in the regulation of PI3K-AKT signaling pathway. CONCLUSIONS:Clarification of the lincRNA genomic clusters specific roles in human cancers could be beneficial for understanding the molecular pathogenesis of different cancer types.

Project description:Multimorbidity (multiple coexisting chronic health conditions) is common and increasing worldwide, and makes care challenging for both patients and healthcare systems. To ensure care is patient-centred rather than specialty-centred, it is important to know which conditions commonly occur together and identify the corresponding patient profile. To date, no studies have described multimorbidity clusters within an unselected hospital population. Our aim was to identify and characterise multimorbidity clusters, in a large, unselected hospitalised patient population. Linked inpatient hospital episode data were used to identify adults admitted to hospital in Grampian, Scotland in 2014 who had ≥ 2 of 30 chronic conditions diagnosed in the 5 years prior. Cluster analysis (Gower distance and Partitioning around Medoids) was used to identify groups of patients with similar conditions. Clusters of conditions were defined based on clinical review and assessment of prevalence within patient groups and labelled according to the most prevalent condition. Patient profiles for each group were described by age, sex, admission type, deprivation and urban-rural area of residence. 11,389 of 41,545 hospitalised patients (27%) had ≥ 2 conditions. Ten clusters of conditions were identified: hypertension; asthma; alcohol misuse; chronic kidney disease and diabetes; chronic kidney disease; chronic pain; cancer; chronic heart failure; diabetes; hypothyroidism. Age ranged from 51 (alcohol misuse) to 79 (chronic heart failure). Women were a higher proportion in the chronic pain and hypothyroidism clusters. The proportion of patients from the most deprived quintile of the population ranged from 6% (hypertension) to 14% (alcohol misuse). Identifying clusters of conditions in hospital patients is a first step towards identifying opportunities to target patient-centred care towards people with unmet needs, leading to improved outcomes and increased efficiency. Here we have demonstrated the face validity of cluster analysis as an exploratory method for identifying clusters of conditions in hospitalised patients with multimorbidity.

Project description:MotivationGene clustering is a widely used technique that has enabled computational prediction of unknown gene functions within a species. However, it remains a challenge to refine gene function prediction by leveraging evolutionarily conserved genes in another species. This challenge calls for a new computational algorithm to identify gene co-clusters in two species, so that genes in each co-cluster exhibit similar expression levels in each species and strong conservation between the species.ResultsHere, we develop the bipartite tight spectral clustering (BiTSC) algorithm, which identifies gene co-clusters in two species based on gene orthology information and gene expression data. BiTSC novelly implements a formulation that encodes gene orthology as a bipartite network and gene expression data as node covariates. This formulation allows BiTSC to adopt and combine the advantages of multiple unsupervised learning techniques: kernel enhancement, bipartite spectral clustering, consensus clustering, tight clustering and hierarchical clustering. As a result, BiTSC is a flexible and robust algorithm capable of identifying informative gene co-clusters without forcing all genes into co-clusters. Another advantage of BiTSC is that it does not rely on any distributional assumptions. Beyond cross-species gene co-clustering, BiTSC also has wide applications as a general algorithm for identifying tight node co-clusters in any bipartite network with node covariates. We demonstrate the accuracy and robustness of BiTSC through comprehensive simulation studies. In a real data example, we use BiTSC to identify conserved gene co-clusters of Drosophila melanogaster and Caenorhabditis elegans, and we perform a series of downstream analysis to both validate BiTSC and verify the biological significance of the identified co-clusters.Availability and implementationThe Python package BiTSC is open-access and available at https://github.com/edensunyidan/BiTSC.Supplementary informationSupplementary data are available at Bioinformatics online.

Dataset Information

Identifying and characterizing high-risk clusters in a heterogeneous ICU population with deep embedded clustering.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets