Dataset Information

A stacking ensemble machine learning model to predict alpha-1 antitrypsin deficiency-associated liver disease clinical outcomes based on UK Biobank data.

ABSTRACT: Alpha-1 antitrypsin deficiency associated liver disease (AATD-LD) is a rare genetic disorder and not well-recognized. Predicting the clinical outcomes of AATD-LD and defining patients more likely to progress to advanced liver disease are crucial for better understanding AATD-LD progression and promoting timely medical intervention. We aimed to develop a tailored machine learning (ML) model to predict the disease progression of AATD-LD. This analysis was conducted through a stacking ensemble learning model by combining five different ML algorithms with 58 predictor variables using nested five-fold cross-validation with repetitions based on the UK Biobank data. Performance of the model was assessed through prediction accuracy, area under the receiver operating characteristic (AUROC), and area under the precision-recall curve (AUPRC). The importance of predictor contributions was evaluated through a feature importance permutation method. The proposed stacking ensemble ML model showed clinically meaningful accuracy and appeared superior to any single ML algorithms in the ensemble, e.g., the AUROC for AATD-LD was 68.1%, 75.9%, 91.2%, and 67.7% for all-cause mortality, liver-related death, liver transplant, and all-cause mortality or liver transplant, respectively. This work supports the use of ML to address the unanswered clinical questions with clinically meaningful accuracy using real-world data.

SUBMITTER: Meng L

PROVIDER: S-EPMC9554039 | biostudies-literature | 2022 Oct

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

A stacking ensemble machine learning model to predict alpha-1 antitrypsin deficiency-associated liver disease clinical outcomes based on UK Biobank data.

Meng Linxi L Treem Will W Heap Graham A GA Chen Jingjing J

Scientific reports 20221011 1

Alpha-1 antitrypsin deficiency associated liver disease (AATD-LD) is a rare genetic disorder and not well-recognized. Predicting the clinical outcomes of AATD-LD and defining patients more likely to progress to advanced liver disease are crucial for better understanding AATD-LD progression and promoting timely medical intervention. We aimed to develop a tailored machine learning (ML) model to predict the disease progression of AATD-LD. This analysis was conducted through a stacking ensemble lear ...[more]

PMID: 36220873

Similar Datasets

Project description:BACKGROUND:Alpha-1-antitrypsin (AAT) deficiency (AATD) of Z, Mmalton, Siiyama type is associated with liver storage of the mutant proteins and liver disease. The Z variant can be diagnosed on isoelectric focusing (IEF) while Mmalton and Siiyama may be missed or misdiagnosed with this technique. Therefore, molecular analysis is mandatory for their characterization. In particular, that holds true for the Mmalton variant as on IEF profile it resembles the wild M2 subtype. METHODS:This is a retrospective analysis involving review of medical records and of liver biopsy specimens from a series of Mmalton, Z and Siiyama Alpha-1-antitrypsin deficiency patients. The review has been implemented by additional histological stains, electron microscopic observations and 3-D modeling studies of the sites of the mutations. RESULTS:Z, Mmalton and Siiyama liver specimen contained characteristic intrahepatocytic PAS-D globules. The globules differed in the three variants as only Mmalton cases showed dark basophilic precipitates within the AAT inclusions. The precipitates were visualized in haematoxylin-eosin (H.E.) stained preparations and corresponded to calcium precipitates as demonstrated by von Kossa staining. On immunohistochemistry, ZAAT inclusions were stained by polyclonal as well as monoclonal noncommercial anti-AAT antibody (AZT11), whilst Mmalton and Siiyama inclusion bodies remained negative with the monoclonal anti-Z antibody. 3-D protein analysis allowed to predict more severe misfolding of the Mmalton molecule as compared to Z and Siiyama that could trigger anomalous interaction with endoplasmic reticulum chaperon proteins, namely calcium binding proteins. CONCLUSIONS:Mmalton AAT inclusion bodies contain calcium precipitates inside them that allow the differential diagnosis with Siiyama and ZAAT inclusions in routine histological sections. The study has confirmed the specificity of the monoclonal AZT11 for the Z mutant. Thus, the combination of these two features is crucial for the distinction between the three variants and for predicting the genotype, whose confirmation would definitely require molecular analysis. Our study provides new data on the pathomorphogenesis of Mmalton inclusion bodies whose mineralization could play a central role in disease pathogenesis of Mmalton that is distinct from the Z and Siiyama variants. Calcium is known to be a major effector of cell death either via the increased intracellular concentration or the alteration of homeostasis.

Project description:Background: Alpha-1 antitrypsin deficiency (A1ATD) is a progressive lung disease caused by inherited pathogenic variants in the SERPINA1 gene. However, their actual role in maintenance of structural and functional characteristics of the corresponding α-1 anti-trypsin (A1AT) protein is not well characterized. Methods: The A1ATD causative SERPINA1 missense variants were initially collected from variant databases, and they were filtered based on their pathogenicity potential. Then, the tertiary protein models were constructed and the impact of individual variants on secondary structure, stability, protein-protein interactions, and molecular dynamic (MD) features of the A1AT protein was studied using diverse computational methods. Results: We identified that A1ATD linked SERPINA1 missense variants like F76S, S77F, L278P, E288V, G216C, and H358R are highly deleterious as per the consensual prediction scores of SIFT, PolyPhen, FATHMM, M-CAP and REVEL computational methods. All these variants were predicted to alter free energy dynamics and destabilize the A1AT protein. These variants were seen to cause minor structural drifts at residue level (RMSD = <2Å) of the protein. Interestingly, S77F and L278P variants subtly alter the size of secondary structural elements like beta pleated sheets and loops. The residue level fluctuations at 100 ns simulation confirm the highly damaging structural consequences of all the six missense variants on the conformation dynamics of the A1AT protein. Moreover, these variants were also predicted to cause functional deformities by negatively impacting the binding energy of A1AT protein with NE ligand molecule. Conclusion: This study adds a new computational biology dimension to interpret the genotype-protein phenotype relationship between SERPINA1 pathogenic variants with its structural plasticity and functional behavior with NE ligand molecule contributing to the Alpha-1-antitrypsin deficiency. Our results support that A1ATD complications correlates with the conformational flexibility and its propensity of A1AT protein polymerization when misfolded.

Dataset Information

A stacking ensemble machine learning model to predict alpha-1 antitrypsin deficiency-associated liver disease clinical outcomes based on UK Biobank data.

Publications

A stacking ensemble machine learning model to predict alpha-1 antitrypsin deficiency-associated liver disease clinical outcomes based on UK Biobank data.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets