Foley Desmond H, Klein Terry A, Kim Heung Chul, Brown Tracy, Wilkerson Richard C, Rueda Leopoldo M
Division of Entomology, Walter Reed Army Institute of Research, 503 Robert Grant Avenue, Silver Spring, MD 20910, USA.
J Am Mosq Control Assoc. 2010 Jun;26(2):210-3. doi: 10.2987/09-5939.1.
Data on molecularly identified adult and larval mosquitoes collected from 104 sites from the Republic of Korea (ROK) in 2007 were used to test the predictive ability of recently reported ecological niche models (ENMs) for 8 potential malaria vectors. The ENMs, based on the program Maxent and the least presence threshold criterion, predicted 100% of new collection locations for Anopheles sinensis, An. belenrae, An. pullus, and An. sineroides; 96% of locations for An. kleini; and 83% for An. lesteri, but were relatively unsuccessful for the infrequently collected non-Hyrcanus group species An. koreicus and An. lindesayi japonicas. The ENMs produced with the use of Maxent had fewer omission errors than those using the Genetic Algorithm for Rule-Set Prediction program. The results emphasize the importance of independent test data for validation and improvement of ENMs, and lend support for the further development of ENMs for predicting the distribution of malaria vectors in the ROK.
2007年从大韩民国104个地点收集的经分子鉴定的成年和幼虫蚊子数据,用于测试最近报告的生态位模型(ENM)对8种潜在疟疾媒介的预测能力。基于Maxent程序和最低出现阈值标准的ENM,对中华按蚊、贝氏按蚊、微小按蚊和中华类按蚊的新采集地点预测准确率为100%;对克莱尼按蚊的地点预测准确率为96%;对雷氏按蚊的预测准确率为83%,但对于不常采集的非赫坎按蚊类物种朝鲜按蚊和日本林氏按蚊,预测相对不成功。使用Maxent生成的ENM比使用规则集预测遗传算法程序产生的遗漏误差更少。结果强调了独立测试数据对验证和改进ENM的重要性,并支持进一步开发ENM以预测大韩民国疟疾媒介的分布。