Objective The Montreal Cognitive Assessment (MoCA) is recommended for assessing general cognition in Parkinson’s disease (PD). Several cutoffs of MoCA scores for diagnosing PD with cognitive impairment (PD-CI) have been proposed, with varying sensitivity and specificity. This study investigated the utility of machine learning algorithms using MoCA cognitive domain scores for improving diagnostic performance for PD-CI.Methods In total, 2,069 MoCA results were obtained from 397 patients with PD enrolled in the Parkinson’s Progression Markers Initiative database with a diagnosis of cognitive status based on comprehensive neuropsychological assessments. Using the same number of MoCA results randomly sampled from patients with PD with normal cognition or PD-CI, discriminant validity was compared between machine learning (logistic regression, support vector machine, or random forest) with domain scores and a cutoff method.Results Based on cognitive status classification using a dataset that permitted sampling of MoCA results from the same individual (n = 221 per group), no difference was observed in accuracy between the cutoff value method (0.74 ± 0.03) and machine learning (0.78 ± 0.03). Using a more stringent dataset that excluded MoCA results (n = 101 per group) from the same patients, the accuracy of the cutoff method (0.66 ± 0.05), but not that of machine learning (0.74 ± 0.07), was significantly reduced. Inclusion of cognitive complaints as an additional variable improved the accuracy of classification using the machine learning method (0.87–0.89).Conclusion Machine learning analysis using MoCA domain scores is a valid method for screening cognitive impairment in PD.

Genomic analysis begins with de novo assembly of short-read fragments in order to reconstruct full-length base sequences without exploiting a reference genome sequence. Then, in the annotation step, gene locations are identified within the base sequences, and the structures and functions of these genes are determined. Recently, a wide range of powerful tools have been developed and published for whole-genome analysis, enabling even individual researchers in small laboratories to perform whole-genome analyses on their objects of interest. However, these analytical tools are generally complex and use diverse algorithms, parameter setting methods, and input formats; thus, it remains difficult for individual researchers to select, utilize, and combine these tools to obtain their final results. To resolve these issues, we have developed a genome analysis pipeline (GAAP) for semiautomated, iterative, and high-throughput analysis of whole-genome data. This pipeline is designed to perform read correction, de novo genome (transcriptome) assembly, gene prediction, and functional annotation using a range of proven tools and databases. We aim to assist non-IT researchers by describing each stage of analysis in detail and discussing current approaches. We also provide practical advice on how to access and use the bioinformatics tools and databases and how to implement the provided suggestions. Whole-genome analysis of Toxocara canis is used as case study to show intermediate results at each stage, demonstrating the practicality of the proposed method.

Kim YJ, et al. Yonsei Med J. 2022 Aug;63(8):724-734. https://doi.org/10.3349/ymj.2022.63.8.724

Objective The Montreal Cognitive Assessment (MoCA) is recommended for general cognitive evaluation in Parkinson’s disease (PD) patients. However, age- and education-adjusted cutoffs specifically for PD have not been developed or systematically validated across PD cohorts with diverse education levels.Methods In this retrospective analysis, we utilized data from 1,293 Korean patients with PD whose cognitive diagnoses were determined through comprehensive neuropsychological assessments. Age- and education-adjusted cutoffs were formulated based on 1,202 patients with PD. To identify the optimal machine learning model, clinical parameters and MoCA domain scores from 416 patients with PD were used. Comparative analyses between machine learning methods and different cutoff criteria were conducted on an additional 91 consecutive patients with PD.Results The cutoffs for cognitive impairment decrease with increasing age within the same education level. Similarly, lower education levels within the same age group correspond to lower cutoffs. For individuals aged 60–80 years, cutoffs were set as follows: 25 or 24 years for those with more than 12 years of education, 23 or 22 years for 10–12 years, and 21 or 20 years for 7–9 years. Comparisons between age- and education-adjusted cutoffs and the machine learning method showed comparable accuracies. The cutoff method resulted in a higher sensitivity (0.8627), whereas machine learning yielded higher specificity (0.8250).Conclusion Both the age- and education-adjusted cutoff methods and machine learning methods demonstrated high effectiveness in detecting cognitive impairment in PD patients. This study highlights the necessity of tailored cutoffs and suggests the potential of machine learning to improve cognitive assessment in PD patients.

Precision medicine has been revolutionized by the advent of high-throughput next-generation sequencing (NGS) technology and development of various bioinformatic analysis tools for large-scale NGS big data. At the population level, biomedical studies have identified human diseases and phenotype-associated genetic variations using NGS technology, such as whole-genome sequencing, exome sequencing, and gene panel sequencing. Furthermore, patients’ genetic variations related to a specific phenotype can also be identified by analyzing their genomic information. These breakthroughs paved the way for the clinical diagnosis and precise treatment of patients’ diseases. Although many bioinformatics tools have been developed to analyze the genetic variations from the individual patient’s NGS data, it is still challenging to develop user-friendly programs for clinical physicians who do not have bioinformatics programing skills to diagnose a patient’s disease using the genomic data. In response to this demand, we developed a Phenotype to Genotype Variation program (PhenGenVar), which is a user-friendly interface for monitoring the variations in a gene of interest for molecular diagnosis. This allows for flexible filtering and browsing of variants of the disease and phenotype-associated genes. To test this program, we analyzed the whole-genome sequencing data of an anonymous person from the 1000 human genome project data. As a result, we were able to identify several genomic variations, including single-nucleotide polymorphism, insertions, and deletions in specific gene regions. Therefore, PhenGenVar can be used to diagnose a patient’s disease. PhenGenVar is freely accessible and is available at our website.