Dataset dependency of low-density lipoprotein-cholesterol estimation by machine learning.

OBJECTIVES: We evaluated the applicability of a machine learning-based low-density lipoprotein-cholesterol (LDL-C) estimation method and the influence of the characteristics of the training datasets. METHODS: Three training datasets were chosen from training datasets: health check-up participants at the Resource Center for Health Science (N = 2664), clinical patients at Gifu University Hospital (N = 7409), and clinical patients at Fujita Health University Hospital (N = 14,842). Nine different machine learning models were constructed through hyperparameter tuning and 10-fold cross-validation. Another test dataset of another 3711 clinical patients at Fujita Health University Hospital was selected as the test set used for comparing and validating the model against the Friedewald formula and the Martin method. RESULTS: The coefficients of determination of the models trained on the health check-up dataset produced coefficients of determination that were equal to or inferior to those of the Martin method. In contrast, the coefficients of determination of several models trained on clinical patients exceeded those of the Martin method. The means of the differences and the convergences to the direct method were higher for the models trained on the clinical patients' dataset than for those trained on the health check-up participants' dataset. The models trained on the latter dataset tended to overestimate the 2019 ESC/EAS Guideline for LDL-cholesterol classification. CONCLUSION: Although machine learning models provide valuable method for LDL-C estimates, they should be trained on datasets with matched characteristics. The versatility of machine learning methods is another important consideration.

Validation of the Martin method to estimate low-density lipoprotein cholesterol concentrations in Japanese populations and a modified method for laboratory information system application.

OBJECTIVES: High concentrations of low-density lipoprotein cholesterol (LDL-C) are a risk factor for cardiovascular disease. We validated the efficacy of the Martin method is useful in the estimation of LDL-C concentrations was validated in Japanese populations and derived a modified Martin method for easy laboratory information system applications. METHODS: We created 3 subject groups, including 2664 health check-up participants registered with the Resource Center for Health Science, 29,806 clinical patients (A) in the Gifu University Hospital, and 113,716 clinical patients (B) in the Fujita Health University Hospital. Each method to estimate serum LDL-C concentrations (Friedewald formula, Martin method and modified Martin method) was validated by correlation analysis with serum LDL-C concentrations measured using a direct method. RESULTS: The correlation coefficients with the direct method in terms of the Friedewald formula, Martin method, and modified Martin method were 0.963, 0.972 and 0.970 in the health check-up participants; 0.946, 0.962 and 0.961 in clinical patients A; and 0.961, 0.979 and 0.978 in clinical patients B, respectively. Concordance ratios with using the direct method in the Friedewald formula, Martin method and modified Martin method were 82.8%, 85.5% and 85.3% in the health check-up participants; 76.4%, 80.5% and 80.2% in clinical patients A; and 76.1%, 82.6% and 82.6% in clinical patients B, respectively. CONCLUSION: Our results show that the Martin and modified Martin methods display good performance in terms of the estimation of LDL-C concentrations among triglyceride concentrations of a wide range, and they may thus be useful for estimating LDL-C concentrations.

Influence of increased amount of exercise on improvements in walking ability of convalescent patients with post-stroke hemiplegia.

[Purpose] This study aimed to determine the effects of increased amount of physical therapy exercise on improvements in the walking ability of patients with stroke. [Subjects and Methods] The subjects were selected from patients with stroke who were hospitalized in the convalescent rehabilitation ward, and included 91 patients who received physical therapy for 2.5-3 exercise sessions per day during 2005-2006 (PT3unit group), and 86 patients who received physical therapy for 4.5-6 exercise sessions per day during 2010-2015 (PT6unit group). The functional independence measure (FIM) score evaluates the walking ability of patients during hospital admission, 2 and 4 weeks after admission, and at discharge. The FIM score was stratified according to the degree of lower limb motor paralysis and subsequently compared between groups. [Results] Among the patients with complete paralysis and severe paralysis, the FIM-Walking scores at 4 weeks after admission and at discharge were significantly higher in the PT6unit group than in the PT3unit group. No significant differences were found between the PT6unit and PT3unit groups for patients with mild and moderate paralysis. [Conclusion] Higher amounts of physical therapy exercise contributed to improvements in the walking ability of patients with complete and severe lower limb paralysis.

Molecular cloning and expression analysis of feline α1-microglobulin.

Full-length cDNA that encodes feline α1-microglobulin (Feα1m)-bikunin was obtained from a feline liver and cloned using an oligo-capping method. The Feα1m-bikunin cDNA was found to contain 1284 nucleotides, and Feα1m was found to include an open reading frame encoding a polypeptide of 201 amino acids. The deduced amino acid sequence of Feα1m showed varying amino acid identity when compared with the published sequences of the related α1-m of other species, ranging from 71.1 to 82.1%. Feα1m mRNA expression was confirmed by reverse transcription polymerase chain reaction (RT-PCR) and real-time PCR analysis in the cerebrum, cerebellum, lung, heart, liver, spleen, pancreas, kidney, adrenal gland, and testicle. The highest Feα1m mRNA level was found in the liver.

Livistona palms in Australia: ancient relics or opportunistic immigrants?

Eighteen of the 34 species of the fan palm genus Livistona (Arecaceae) are restricted to Australia and southern New Guinea, east of Wallace's Line, an ancient biogeographic boundary between the former supercontinents Laurasia and Gondwana. The remaining species extend from SE Asia to Africa, west of Wallace's Line. Competing hypotheses contend that Livistona is (a) ancient, its current distribution a relict of the supercontinents, or (b) a Miocene immigrant from the north into Australia as it drifted towards Asia. We have tested these hypotheses using Bayesian and penalized likelihood molecular dating based on 4Kb of nuclear and chloroplast DNA sequences with multiple fossil calibration points. Ancestral areas and biomes were reconstructed using parsimony and maximum likelihood. We found strong support for the second hypothesis, that a single Livistona ancestor colonized Australia from the north about 10-17Ma. Spread and diversification of the genus within Australia was likely favoured by a transition from the aseasonal wet to monsoonal biome, to which it could have been preadapted by fire-tolerance.