Predicting mortality risk for preterm infants using random forest.
Ontology highlight
ABSTRACT:
SUBMITTER: Lee J
PROVIDER: S-EPMC8012581 | biostudies-literature |
REPOSITORIES: biostudies-literature
Similar Datasets
Project description:Mortality remains an exceptional burden of extremely preterm birth. Current clinical mortality prediction scores are calculated using a few static variable measurements, such as gestational age, birth weight, temperature, and blood pressure at admission. While these models do provide some insight, numerical and time-series vital sign data are also available for preterm babies admitted to the NICU and may provide greater insight into outcomes. Computational models that predict the mortality risk of preterm birth in the NICU by integrating vital sign data and static clinical variables in real time may be clinically helpful and potentially superior to static prediction models. However, there is a lack of established computational models for this specific task. In this study, we developed a novel deep learning model, DeepPBSMonitor (Deep Preterm Birth Survival Risk Monitor), to predict the mortality risk of preterm infants during initial NICU hospitalization. The proposed deep learning model can effectively integrate time-series vital sign data and fixed variables while resolving the influence of noise and imbalanced data. The proposed model was evaluated and compared with other approaches using data from 285 infants. Results showed that the DeepPBSMonitor model outperforms other approaches, with an accuracy, recall, and AUC score of 0.888, 0.780, and 0.897, respectively. In conclusion, the proposed model has demonstrated efficacy in predicting the real-time mortality risk of preterm infants in initial NICU hospitalization.
| S-EPMC8280207 | biostudies-literature
Project description:Background: Understanding the location and cell-type specific binding of Transcription Factors (TFs) is important in the study of gene regulation. Computational prediction of TF binding sites is challenging, because TFs often bind only to short DNA motifs and cell-type specific co-factors may work together with the same TF to determine binding. Here, we consider the problem of learning a general model for the prediction of TF binding using DNase1-seq data and TF motif description in form of position specific energy matrices (PSEMs). Methods: We use TF ChIP-seq data as a gold-standard for model training and evaluation. Our contribution is a novel ensemble learning approach using random forest classifiers. In the context of the ENCODE-DREAM in vivo TF binding site prediction challenge we consider different learning setups. Results: Our results indicate that the ensemble learning approach is able to better generalize across tissues and cell-types compared to individual tissue-specific classifiers or a classifier built based upon data aggregated across tissues. Furthermore, we show that incorporating DNase1-seq peaks is essential to reduce the false positive rate of TF binding predictions compared to considering the raw DNase1 signal. Conclusions: Analysis of important features reveals that the models preferentially select motifs of other TFs that are close interaction partners in existing protein protein-interaction networks. Code generated in the scope of this project is available on GitHub: https://github.com/SchulzLab/TFAnalysis (DOI: 10.5281/zenodo.1409697).
| S-EPMC6823902 | biostudies-literature
Project description:Psoriasis is a chronic inflammatory skin disease that affects approximately 125 million people worldwide. It has significant impacts on both physical and emotional health-related quality of life comparable to other major illnesses. Accurately prediction of psoriasis using biomarkers from routine laboratory tests has important practical values. Our goal is to derive a powerful predictive model for psoriasis disease based on only routine hospital tests. We collected a data set including 466 psoriasis patients and 520 healthy controls with 81 variables from only laboratory routine tests, such as age, total cholesterol, HDL cholesterol, blood pressure, albumin, and platelet distribution width. In this study, Boruta feature selection method was applied to select the most relevant features, with which a Random Forest model was constructed. The model was tested with 30 repetitions of 10-fold cross-validation. Our classification model yielded an average accuracy of 86.9%. 26 notable features were selected by Boruta, among which 15 features are confirmed from previous studies, and the rest are worth further investigations. The experimental results demonstrate that the machine learning approach has good potential in predictive modeling for the psoriasis disease given the information only from routine hospital tests.
| S-EPMC8525763 | biostudies-literature
Project description:Influenza A viruses remain a significant health problem, especially when a novel subtype emerges from the avian population to cause severe outbreaks in humans. Zoonotic viruses arise from the animal population as a result of mutations and reassortments, giving rise to novel strains with the capability to evade the host species barrier and cause human infections. Despite progress in understanding interspecies transmission of influenza viruses, we are no closer to predicting zoonotic strains that can lead to an outbreak. We have previously discovered distinct host tropism protein signatures of avian, human and zoonotic influenza strains obtained from host tropism predictions on individual protein sequences. Here, we apply machine learning approaches on the signatures to build a computational model capable of predicting zoonotic strains. The zoonotic strain prediction model can classify avian, human or zoonotic strains with high accuracy, as well as providing an estimated zoonotic risk. This would therefore allow us to quickly determine if an influenza virus strain has the potential to be zoonotic using only protein sequences. The swift identification of potential zoonotic strains in the animal population using the zoonotic strain prediction model could provide us with an early indication of an imminent influenza outbreak.
| S-EPMC5485959 | biostudies-literature
Project description:BackgroundWe present a method utilizing Healthcare Cost and Utilization Project (HCUP) dataset for predicting disease risk of individuals based on their medical diagnosis history. The presented methodology may be incorporated in a variety of applications such as risk management, tailored health communication and decision support systems in healthcare.MethodsWe employed the National Inpatient Sample (NIS) data, which is publicly available through Healthcare Cost and Utilization Project (HCUP), to train random forest classifiers for disease prediction. Since the HCUP data is highly imbalanced, we employed an ensemble learning approach based on repeated random sub-sampling. This technique divides the training data into multiple sub-samples, while ensuring that each sub-sample is fully balanced. We compared the performance of support vector machine (SVM), bagging, boosting and RF to predict the risk of eight chronic diseases.ResultsWe predicted eight disease categories. Overall, the RF ensemble learning method outperformed SVM, bagging and boosting in terms of the area under the receiver operating characteristic (ROC) curve (AUC). In addition, RF has the advantage of computing the importance of each variable in the classification process.ConclusionsIn combining repeated random sub-sampling with RF, we were able to overcome the class imbalance problem and achieve promising results. Using the national HCUP data set, we predicted eight disease categories with an average AUC of 88.79%.
| S-EPMC3163175 | biostudies-literature
Project description:BackgroundMajority of influenza A viruses reside and circulate among animal populations, seldom infecting humans due to host range restriction. Yet when some avian strains do acquire the ability to overcome species barrier, they might become adapted to humans, replicating efficiently and causing diseases, leading to potential pandemic. With the huge influenza A virus reservoir in wild birds, it is a cause for concern when a new influenza strain emerges with the ability to cross host species barrier, as shown in light of the recent H7N9 outbreak in China. Several influenza proteins have been shown to be major determinants in host tropism. Further understanding and determining host tropism would be important in identifying zoonotic influenza virus strains capable of crossing species barrier and infecting humans.ResultsIn this study, computational models for 11 influenza proteins have been constructed using the machine learning algorithm random forest for prediction of host tropism. The prediction models were trained on influenza protein sequences isolated from both avian and human samples, which were transformed into amino acid physicochemical properties feature vectors. The results were highly accurate prediction models (ACC>96.57; AUC>0.980; MCC>0.916) capable of determining host tropism of individual influenza proteins. In addition, features from all 11 proteins were used to construct a combined model to predict host tropism of influenza virus strains. This would help assess a novel influenza strain's host range capability.ConclusionsFrom the prediction models constructed, all achieved high prediction performance, indicating clear distinctions in both avian and human proteins. When used together as a host tropism prediction system, zoonotic strains could potentially be identified based on different protein prediction results. Understanding and predicting host tropism of influenza proteins lay an important foundation for future work in constructing computation models capable of directly predicting interspecies transmission of influenza viruses. The models are available for prediction at http://fluleap.bic.nus.edu.sg.
| S-EPMC4290784 | biostudies-literature
Project description:BackgroundThis study aimed to compare the predictive values of eight scoring systems (Neonatal Critical Illness Score [NCIS], Neonatal Therapeutical Intervention Score System [NTISS], Clinical Risk Index for Babies [CRIB], Clinical Risk Index for Babies II [CRIB-II], Score for Neonatal Acute Physiology Perinatal Extension [SNAPPE], Score for Neonatal Acute Physiology Perinatal Extension II [SNAPPE-II], Score for Neonatal Acute Physiology [SNAP], and Score for Neonatal Acute Physiology II [SNAP-II]) for the mortality risk among preterm infants.MethodsThe Embase, PubMed, Chinese Biomedical Database, Web of Science, and Cochrane Library databases were searched to collect studies that compared different scoring systems in predicting the mortality risk in preterm infants from database inception to March 2023. Literature screening, data extraction, and bias risk assessment were independently conducted by two researchers. Subsequently, the random-effects model was used for the network meta-analysis.ResultsA total of 19 articles were included, comprising 14,377 preterm infants and 8 scoring systems. Compared to CRIB-II, NCIS, NTISS, SNAP-II, and SNAPPE-II, CRIB demonstrated better predictive efficiency for preterm infant mortality risk (P < 0.05). Relative to CRIB, CRIB-II, and SNAPPE, SNAP-II had worse predictive efficiency for preterm infant mortality risk (P < 0.05). The surface under the cumulative ranking curve of the eight scoring systems was as follows: CRIB (0.980) > SNAPPE (0.718) >SNAP (0.534) >CRIB-II (0.525) >NTISS (0.478) >NCIS (0.422) >SNAPPE-II (0.298) >SNAP-II (0.046).ConclusionThe CRIB scoring system showed the highest accuracy in predicting preterm infant mortality risk and was simple to perform. Therefore, CRIB selection can be prioritized in clinical practice.Systematic review registrationhttps://www.crd.york.ac.uk/prospero/display_record.php?RecordID=434731, PROSPERO (CRD42023434731).
| S-EPMC10757321 | biostudies-literature
Project description:In recent years, glycosaminoglycans (GAGs) have emerged into the focus of biochemical and biomedical research due to their importance in a variety of physiological processes. These molecules show great diversity, which makes their analysis highly challenging. A promising tool for identifying the structural motifs and conformation of shorter GAG chains is cryogenic gas-phase infrared (IR) spectroscopy. In this work, the cryogenic gas-phase IR spectra of mass-selected heparan sulfate (HS) di-, tetra-, and hexasaccharide ions were recorded to extract vibrational features that are characteristic to structural motifs. The data were augmented with chondroitin sulfate (CS) disaccharide spectra to assemble a training library for random forest (RF) classifiers. These were used to discriminate between GAG classes (CS or HS) and different sulfate positions (2-O-, 4-O-, 6-O-, and N-sulfation). With optimized data preprocessing and RF modeling, a prediction accuracy of >97% was achieved for HS tetra- and hexasaccharides based on a training set of only 21 spectra. These results exemplify the importance of combining gas-phase cryogenic IR ion spectroscopy with machine learning to improve the future analytical workflow for GAG sequencing and that of other biomolecules, such as metabolites.
| S-EPMC10103134 | biostudies-literature
Project description:BackgroundAssessing the risk of developing clinical Alzheimer's disease (AD) dementia, by machine learning survival analysis approaches, among participants registered in Alzheimer's Disease Centers is important for AD dementia management.ObjectiveTo construct a prediction model for the onset time of clinical AD dementia using the National Alzheimer Coordinating Center (NACC) and the Alzheimer's Disease Neuroimaging Initiative (ADNI) registered cohorts.MethodsA model was constructed using the Random Survival Forest (RSF) approach and internally and externally validated on the NACC cohort and the ADNI cohort. An R package and a Shiny app were provided for accessing the model.ResultsWe built a predictive model having the six predictors: delayed logical memory score (story recall), CDR® Dementia Staging Instrument - Sum of Boxes, general orientation in CDR®, ability to remember dates and ability to pay bills in the Functional Activities Questionnaire, and patient age. The C indices of the model were 90.82% (SE = 0.71%) and 86.51% (SE = 0.75%) in NACC and ADNI respectively. The time-dependent AUC and accuracy at 48 months were 92.48% (SE = 1.12%) and 88.66% (SE = 1.00%) respectively in NACC, and 90.16% (SE = 1.12%) and 85.00% (SE = 1.14%) respectively in ADNI.ConclusionThe model showed good prediction performance and the six predictors were easy to obtain, cost-effective, and non-invasive. The model could be used to inform clinicians and patients on the probability of developing clinical AD dementia in 4 years with high accuracy.
| S-EPMC10529100 | biostudies-literature
Project description:BackgroundThe long-term prognosis of the cardio-metabolic and renal complications, in addition to mortality in patients with newly diagnosed pulmonary hypertension, are unclear. This study aims to develop a scalable predictive model in the form of an electronic frailty index (eFI) to predict different adverse outcomes.MethodsThis was a population-based cohort study of patients diagnosed with pulmonary hypertension between January 1st, 2000 and December 31st, 2017, in Hong Kong public hospitals. The primary outcomes were mortality, cardiovascular complications, renal diseases, and diabetes mellitus. The univariable and multivariable Cox regression analyses were applied to identify the significant risk factors, which were fed into the non-parametric random survival forest (RSF) model to develop an eFI.ResultsA total of 2,560 patients with a mean age of 63.4 years old (interquartile range: 38.0-79.0) were included. Over a follow-up, 1,347 died and 1,878, 437, and 684 patients developed cardiovascular complications, diabetes mellitus, and renal disease, respectively. The RSF-model-identified age, average readmission, anti-hypertensive drugs, cumulative length of stay, and total bilirubin were among the most important risk factors for predicting mortality. Pair-wise interactions of factors including diagnosis age, average readmission interval, and cumulative hospital stay were also crucial for the mortality prediction. Patients who developed all-cause mortality had higher values of the eFI compared to those who survived (P < 0.0001). An eFI ≥ 9.5 was associated with increased risks of mortality [hazard ratio (HR): 1.90; 95% confidence interval [CI]: 1.70-2.12; P < 0.0001]. The cumulative hazards were higher among patients who were 65 years old or above with eFI ≥ 9.5. Using the same cut-off point, the eFI predicted a long-term mortality over 10 years (HR: 1.71; 95% CI: 1.53-1.90; P < 0.0001). Compared to the multivariable Cox regression, the precision, recall, area under the curve (AUC), and C-index were significantly higher for RSF in the prediction of outcomes.ConclusionThe RSF models identified the novel risk factors and interactions for the development of complications and mortality. The eFI constructed by RSF accurately predicts the complications and mortality of patients with pulmonary hypertension, especially among the elderly.
| S-EPMC9304657 | biostudies-literature