Background and Purpose—: The prediction of long-term outcomes in ischemic stroke patients may be useful in treatment decisions. Machine learning techniques are being increasingly adapted for use in the medical field because of their high accuracy. This study investigated the applicability of machine learning techniques to predict long-term outcomes in ischemic stroke patients. Methods—: This was a retrospective study using a prospective cohort that enrolled patients with acute ischemic stroke. Favorable outcome was defined as modified Rankin Scale score 0, 1, or 2 at 3 months. We developed 3 machine learning models (deep neural network, random forest, and logistic regression) and compared their predictability. To evaluate the accuracy of the machine learning models, we also compared them to the Acute Stroke Registry and Analysis of Lausanne (ASTRAL) score. Results—: A total of 2604 patients were included in this study, and 2043 (78%) of them had favorable outcomes. The area under the curve for the deep neural network model was significantly higher than that of the ASTRAL score (0.888 versus 0.839; P <0.001), while the areas under the curves of the random forest (0.857; P =0.136) and logistic regression (0.849; P =0.413) models were not significantly higher than that of the ASTRAL score. Using only the 6 variables that are used for the ASTRAL score, the performance of the machine learning models did not significantly differ from that of the ASTRAL score. Conclusions—: Machine learning algorithms, particularly the deep neural network, can improve the prediction of long-term outcomes in ischemic stroke patients.

Shin HB, et al. Yonsei Med J. 2015 Mar;56(2):550-555. https://doi.org/10.3349/ymj.2015.56.2.550

Aberrant epithelial bicarbonate (HCO3-) secretion caused by mutations in the cystic fibrosis transmembrane conductance regulator (CFTR) gene is associated with several diseases including cystic fibrosis and pancreatitis. Dynamically regulated ion channel activity and anion selectivity of CFTR by kinases sensitive to intracellular chloride concentration ([Cl-]i) play an important role in epithelial HCO3- secretion. However, the molecular mechanisms of how [Cl-]i-dependent mechanisms regulate CFTR are unknown. We examined the mechanisms of the CFTR HCO3- channel regulation by [Cl-]i-sensitive kinases using an integrated electrophysiological, molecular, and computational approach including whole-cell, outside-out, and inside-out patch clamp recordings and molecular dissection of WNK1 and CFTR proteins. In addition, we analyzed the effects of pancreatitis-causing CFTR mutations on the WNK1-mediated regulation of CFTR. Among the WNK1, SPAK, and OSR1 kinases that constitute a [Cl-]i-sensitive kinase cascade, the expression of WNK1 alone was sufficient to increase the CFTR bicarbonate permeability (PHCO3/PCl) and conductance (GHCO3) in patch clamp recordings. Molecular dissection of the WNK1 domains revealed that the WNK1 kinase domain is responsible for CFTR PHCO3/PCl regulation by direct association with CFTR, while the surrounding N-terminal regions mediate the [Cl-]i-sensitivity of WNK1. Furthermore, the pancreatitis-causing R74Q and R75Q mutations in the elbow helix 1 of CFTR hampered WNK1-CFTR physical associations and reduced WNK1-mediated CFTR PHCO3/PCl regulation. The CFTR HCO3- channel activity is regulated by [Cl-]i and a WNK1-dependent mechanism. Our results provide new insights into the regulation of the ion selectivity of CFTR and the pathogenesis of CFTR-related disorders.

To date, approximately 50 short tandem repeat (STR) disorders have been identified; yet, clinical laboratories rarely conduct STR analysis on exomes. To assess its diagnostic value, we analyzed STRs in 6099 exomes from 2510 families with mostly suspected neurogenetic disorders. We employed ExpansionHunter and REViewer to detect pathogenic repeat expansions, confirming them using orthogonal methods. Genotype-phenotype correlations led to the diagnosis of thirteen individuals in seven previously undiagnosed families, identifying three autosomal dominant disorders: dentatorubral-pallidoluysian atrophy (n = 3), spinocerebellar ataxia type 7 (n = 2), and myotonic dystrophy type 1 (n = 2), resulting in a diagnostic gain of 0.28% (7/2510). Additionally, we found expanded ATXN1 alleles (≥39 repeats) with varying patterns of CAT interruptions in twelve individuals, accounting for approximately 0.19% in the Korean population. Our study underscores the importance of integrating STR analysis into exome sequencing pipeline, broadening the application of exome sequencing for STR assessments.

BACKGROUND: Genetic factors play an important role in the pathogenesis of craniosynostosis (CRS). However, the molecular diagnosis of CRS in clinical practice is limited because of its heterogeneous etiology. OBJECTIVE: To investigate the genomic landscape of CRS in a Korean cohort and also to establish a practical diagnostic workflow by applying targeted panel sequencing. METHODS: We designed a customized panel covering 34 CRS-related genes using in-solution hybrid capture method. We enrolled 110 unrelated Korean patients with CRS, including 40 syndromic and 70 nonsyndromic cases. A diagnostic pipeline was established by combining in-depth clinical reviews and multiple bioinformatics tools for analyzing single-nucleotide variants (SNV)s and copy number variants (CNV)s. RESULTS: The diagnostic yield of the targeted panel was 30.0% (33/110). Twenty-five patients (22.7%) had causal genetic variations resulting from SNVs or indels in 9 target genes (TWIST1, FGFR3, TCF12, ERF, FGFR2, ALPL, EFNB1, FBN1, and SKI, in order of frequency). CNV analysis identified 8 (7.3%) additional patients with chromosomal abnormalities involving 1p32.3p31.3, 7p21.1, 10q26, 15q21.3, 16p11.2, and 17p13.3 regions; these cases mostly presented with syndromic clinical features. CONCLUSION: The present study shows the wide genomic landscape of CRS, revealing various genetic factors for CRS pathogenesis. In addition, the results demonstrate that an efficient diagnostic workup using target panel sequencing provides great clinical utility in the molecular diagnosis of CRS.

GGC repeat expansions in the NOTCH2NLC gene are associated with a broad spectrum of progressive neurologic disorders, notably, neuronal intranuclear inclusion disease (NIID). We aimed to investigate the population-wide prevalence and clinical manifestations of NOTCH2NLC-related disorders in Koreans. We conducted a study using 2 different cohorts from the Korean population. Patients with available brain MRI scans from Seoul National University Hospital (SNUH) were thoroughly reviewed, and NIID-suspected patients presenting the zigzag edging signs underwent genetic evaluation for NOTCH2NLC repeats by Cas9-mediated nanopore sequencing. In addition, we analyzed whole-genome sequencing data from 3,887 individuals in the Korea Biobank cohort to estimate the distribution of the repeat counts in Koreans and to identify putative patients with expanded alleles and neurologic phenotypes. In the SNUH cohort, among 90 adult-onset leukoencephalopathy patients with unknown etiologies, we found 20 patients with zigzag edging signs. Except for 2 diagnosed with fragile X-associated tremor/ataxia syndrome and 2 with unavailable samples, all 16 patients (17.8%) were diagnosed with NIID (repeat range: 87–217). By analyzing the Korea Biobank cohort, we estimated the distribution of repeat counts and threshold (>64) for Koreans, identifying 6 potential patients with NIID. Furthermore, long-read sequencing enabled the elucidation of transmission and epigenetic patterns of NOTCH2NLC repeats within a family affected by pediatric-onset NIID. This study presents the population-wide distribution of NOTCH2NLC repeats and the estimated prevalence of NIID in Koreans, providing valuable insights into the association between repeat counts and disease manifestations in diverse neurologic disorders.