Biomedical Informatics and Cheminformatics Group, Conjugate and Medicinal Chemistry Laboratory, Department of Radiology, Brigham and Women's Hospital, Harvard Medical School, Boston, MA 02115, USA.
IEEE Trans Nanobioscience. 2010 Mar;9(1):44-50. doi: 10.1109/TNB.2009.2037745. Epub 2010 Jan 19.
Clustering is the grouping of similar objects into a class. Local clustering feature refers to the phenomenon whereby one group of data is separated from another, and the data from these different groups are clustered locally. A compact class is defined as one cluster in which all similar elements cluster tightly within the cluster. Herein, the essence of the local clustering feature, revealed by mathematical manipulation, results in a novel clustering algorithm termed as the special local clustering (SLC) algorithm that was used to process gene microarray data related to Alzheimer's disease (AD). SLC algorithm was able to group together genes with similar expression patterns and identify significantly varied gene expression values as isolated points. If a gene belongs to a compact class in control data and appears as an isolated point in incipient, moderate and/or severe AD gene microarray data, this gene is possibly associated with AD. Application of a clustering algorithm in disease-associated gene identification such as in AD is rarely reported.
聚类是将相似对象分组到一个类中。局部聚类特征是指一组数据与另一组数据分开的现象，来自这些不同组的数据在局部被聚类。紧密度是指一个聚类中所有相似的元素在聚类内部紧密聚集。通过数学运算揭示了局部聚类特征的本质，产生了一种新的聚类算法，称为特殊局部聚类（SLC）算法，用于处理与阿尔茨海默病（AD）相关的基因微阵列数据。SLC 算法能够将具有相似表达模式的基因组合在一起，并将显著变化的基因表达值识别为孤立点。如果一个基因在对照数据中属于一个紧凑类，并且在早期、中度和/或重度 AD 基因微阵列数据中表现为一个孤立点，那么这个基因可能与 AD 有关。聚类算法在疾病相关基因识别中的应用，如 AD 中，很少有报道。
