决策树在天津市某区公务员健康状况影响因素分析中的应用

目的了解天津市某区公务员健康状况的影响因素,为提高该人群的健康水平提供依据。方法于2008年9-12月,采用整群抽样的方法抽取天津市某区部分公务员,发放问卷740份,进行健康状况和相关影响因素的问卷调查。利用705份有效问卷,应用SAS8.2 Enterprise Miner模块建立决策树模型,对该区公务员人群的健康状况影响因素进行分析和预测。结果该区公务员总体患病率为47.0%,患病率较高的前4种疾病为高血压(17.87%),高血脂(17.73%),脂肪肝(15.89%),颈、腰椎疾病(12.77%)。决策树筛检出的健康状况影响危险因素(重要性)包括年龄大(1.0000)、吸烟(0.8359)、不按时吃饭(0.7972)、心理健康分值低(0.5860)、体质指数高(0.4748)、被动吸烟(0.3673)、教育系统公务员(0.2876)、文化程度高(0.2832)、饮酒(0.1365);保护性因素(重要性)包括充足的睡眠时间(0.3873)、充足的体育锻炼时间(0.2845)、女性(0.2636)、亚健康分值低(0.2364)。决策树模型ROC曲线下面积为0.8881(95%CI:0.8643～0.9119),预测的准确度为80%。结论公务员人群健康状况不容乐观,各种慢性病患病率较高,是今后开展健康管理的重点群体。     

基于前列腺影像报告和数据系统第2版的机器学习模型诊断高级别前列腺癌

目的 建立基于前列腺影像报告和数据系统第2版（PI-RADS v2）的支持向量机（SVM）、决策树（DT）和Logistic回归3种机器学习模型,评价上述模型对高级别前列腺癌的诊断价值。方法 回顾性分析于我院接受前列腺多参数MR扫描并取得病理结果的194例患者的资料,其中高级别癌63例,非高级别癌131例。将评价因素（PI-RADS v2评分、年龄、游离前列腺特异抗原、前列腺特异性抗原比值、前列腺特异抗原密度）录入SVM、DT和Logistic回归3种机器学习模型进行诊断,通过ROC曲线评价PI-RADS v2评分和3种机器学习模型诊断高级别前列腺癌的价值。结果 PI-RADS v2、SVM、DT和Logistic回归模型诊断高级别前列腺癌的敏感度分别为72.73%、69.09%、87.27%和70.91%;特异度分别为87.29%、93.22%、93.22%和95.76%。DT模型诊断高级别前列腺癌ROC的AUC（AUC=0.90,P<0.01）最大,且与PI-RADS v2评分、SVM、Logistic回归比较差异均有统计学意义（P均<0.05）。结论 PI-RADS v2评分、SVM、DT和Logistic回归模型诊断高级别前列腺癌的价值均较好。     

Antioxidant and antiacetylcholinesterase activities of Sorbus torminalis (L.) Crantz (wild service tree) fruits

In this study, the antioxidant and antiacetylcholinesterase activities of Sorbus torminalis (L.) Crantz fruits were evaluated. Total phenolic and flavonoid compounds, 2,2′-azino-bis (3-ethylbenzothioazoline-6-sulfonic acid), 2,2′-diphenyl-1-picrylhydrazyl, and superoxide anion radicals scavenging activities and ferric-reducing antioxidant power of water, ethyl acetate, acetone, and methanol extracts were determined for the measurement of the antioxidant activity. Quercetin and α-tocopherol were used as standard antioxidants. The inhibitory effect of the water extract on acetylcholinesterase (AChE) was evaluated using the Ellman method and galantamine was used as a standard. Water extract had the highest total phenolic concentration and the strongest antioxidant activity followed by ethyl acetate and acetone extracts whereas methanol extract has the lowest phenolics and weakest antioxidant activity. Moreover, water extract showed moderate ability to inhibit AChE. It was concluded that fruits of S. torminalis have antioxidant and anti-AChE activities and that the plant might be a natural source of antioxidants and AChE inhibitors.     

Solubilización in vitro del aceite de árbol de té y primeros resultados de su efecto antifúngico en onicomicosis

IntroductionOnychomycosis is the main cause of nail alteration. Hepatotoxicity, interference and low adherence to pharmacological treatment are associated. Therefore, our objective was to assess the in vitro effectiveness of tea tree essential oil (less harmful) against main causative agents of these infections.Material and methodsTrichophyton rubrum and Trichophyton mentagrophytes were isolated and inoculated at a concentration of 3x10⁵ CFU / mL in potato agar dextrose and tea tree essential oil at different concentrations to assess its effect by counting colony forming units and radial growth.ResultsTrichophyton rubrum growth inhibition was obtained at concentrations higher than 0.04% of the essential tea tree oil (p = 0.004). In the case of Trichophyton mentagrophytes, inhibition was obtained at 0.02% (p = 0.017), and even complete inhibition at a final concentration of the oil at 0.07%.ConclusionsTea tree essential oil inhibits the in vitro growth of the fungus and may be a less harmful alternative to the onychomycosis treatment.     

Application of the IML Resistograph to the infield assessment of basic density in plantation eucalypts

Average bark-to-bark resistance of the IML Resistograph PD400 (hereafter referred to as ‘Resi’) was found to provide strong linear correlations with the basic density of 12-mm-diameter increment cores taken from standing plantation eucalypt trees. Relationships between Resi values and approximately 2 000 cores (predominantly Eucalyptus globulus but some E. nitens) were examined across seven studies (representing samples from nine distinct sites) in Tasmania, Victoria and Western Australia. Custom-written software was developed to process the Resi traces to automatically perform a linear baseline correction of the trace, and extract:

• over-bark and under-bark diameter

• bark thickness

• average resistance of the bark-to-bark (under-bark) trace

• average resistance of the outer 50 mm on the entry and exit side of the traces.

Baseline correction was needed to counter the variable effects among trees of needle drag across the diameter, largely a function of tree diameter and wood density.

Individual traces were collected in less than 20 s tree^?1. The sampling conditions of 150 cm min^?1 speed of forward movement (feed speed) and 3 500 revolutions per minute (rpm) were identified as optimal for the plantation eucalypts studied. The relationship between different feed speeds and rpm were linear and coefficients determined to allow average Resi resistance to be converted to a common set of sampling conditions. A simple linear regression was identified in each study to define a slope and intercept to convert the Resi values to basic density and determine the variance between them. Resi traces from different studies were not always collected using the same instrument and this is believed to explain most of the between-study variance in slope and intercept.

Trace processing software was built into a web-based package that is available to make trace processing easy, with defined variables downloadable for use in routine plantation assessment. Users can change the default slope and intercept values to suit their individual instruments or species as required. Further work is required to fully define the effects of instrument, site and species on these relationships.

The IML Resistograph PD400 was found to be an accurate and quick (40–70 trees hour^?1 person^?1) infield tool for estimating diameter and wood density in standing trees. When combined with automated, web-based processing software the methodology is among the lowest cost options conceivable for wood density assessment of plantation eucalypts.     

深圳市新分离两株基孔肯雅热病毒的分子遗传学分析

目的 对2010～2012年深圳市报告的两起输入性基孔肯雅热病原体进行分子遗传学分析.方法 利用C6/36细胞从病人血清中分离基孔肯雅病毒(Chikungunya virus,CHIKV),对分离得到的CHIKV进行全基因组测序和构建系统发生树,结合流行病学资料对其分子遗传特征进行分析.结果 成功分离得到两株CHIKV,分别命名为SZ_20101028和SZ 20120702,SZ_20101028全基因组长为12377bp,SZ_20120823全基因组长为11893bp;构建系统发生树分析,结果表明,SZ_20101028是印度洋亚型,ECSA型的后裔.SZ_20120702为Asian亚型.结论 SZ_20101028与最近10年来在印度洋岛屿爆发流行新亚型且与2010年由输入性引发起东莞疫情爆发的毒株亲源性最高,为99％;而SZ_20120702为Asian亚型,不是近年来爆发流行株.