资源描述:
《一种基于主集分割的基因芯片聚类算法》由会员上传分享,免费在线阅读,更多相关内容在学术论文-天天文库。
1、1000-9825/2005/16(09)1591©2005JournalofSoftware软件学报Vol.16,No.9∗一种基于主集分割的基因芯片聚类算法1+212341滕莉,付旭平,李宏宇,李瑶,陈文斌,李荣宇,沈一帆1(复旦大学计算机科学与工程系,上海200433)2(复旦大学生命科学学院遗传研究所,上海200433)3(复旦大学数学系,上海200433)4(上海博星基因芯片有限责任公司,上海200092)AMicroarrayClusterAlgorithmBasedonDominantSetSegmentation1+212341TENGLi,FUXu-P
2、ing,LIHong-Yu,LIYao,CHENWen-Bin,LIRong-Yu,SHENYi-Fan1(DepartmentofComputerScienceandEngineering,FudanUniversity,Shanghai200433,China)2(InstituteofGenetics,SchoolofLifeScience,FudanUniversity,Shanghai200433,China)3(DepartmentofMathematics,FudanUniversity,Shanghai200433,China)4(ShanghaiBio
3、StarGenechipInc.,Shanghai200092,China)+Correspondingauthor:Phn:+86-852-60801741,E-mail:tengli.hust@263.net,http://www.cse.cuhk.edu.hk/~lteng/Received2004-05-31;Accepted2005-02-04TengL,FuXP,LiHY,LiY,ChenWB,LiRY,ShenYF.Amicroarrayclusteralgorithmbasedondominantsetsegmentation.JournalofSoft
4、ware,2005,16(9):1591−1598.DOI:10.1360/jos161591Abstract:Clusteringalgorithmsarewildlyusedintheresearchofmicroarraydatatoextractgroupsofgenesorsamplesthataretightlycoexpressed.Inmostofthem,someparametersshouldbepredefinedartificially,however,itisverydifficulttodeterminethemmanuallywithout
5、priordomainknowledge.Tohandlethisproblem,aniterativeclusteringalgorithmisproposed.Firstly,bysortingtheoriginaldatabydominantset,similargeneswouldbealignedtogether.It’shardtospecifytheclusterboundary.Acriterionispresentedtopartitionaclusterfromthesorteddataaccordingtothepropertythatthedis
6、tancesbetweentheinsideelementsaresmallerthanthatofoutsideelements.Theideaistoremovetheclusterformthecurrentdataset,repeattheprocess,andstopthealgorithmwhenthestopcriterionsaresatisfied.Thenewclusteringalgorithmisanalyzedonseveralaspectsandtestedonthepublishedyeastcell-cyclemicroarraydata
7、.Theresultsoftheapplicationconfirmthatthemethodisveryapplicable,efficientandhasgoodabilitytoresistnoise.Keywords:microarray;dominantset;clustering;coexpressed;sorting摘要:聚类算法广泛应用于生物芯片数据分析中,用于寻找表达相似的基因或样本.大多数已有算法都需要∗SupportedbytheNationalNaturalScienceFoundationofChinaunder