Efficient Gene Expression Analysis by Linking Multiple Data Mining Algorithms

Abstract

The set of gene micro-arrays, which consists of two leukemia types, was used as a target to evaluate the efficiency of novel integrated data mining classification process. Discovering the most relevant subset of genes among few thousands of analyzed genes is necessary to get accurate disease classification. Dimensional complexity of the classification process was reduced by a filter based on mutual information feature selection coupled with the support vector machines classifier in the leave-one-out loop. The result was an efficient and reliable tool named MIFS/SVM hybrid. Optimal procedure parameters that enable accurate classification and attribute selection could be determined within an acceptable time frame.

K-Means Clustering on 3^rd order polynomial based normalization of Acute Myeloid Leukemia (AML) and Acute Lymphocyte Leukemia (ALL)

“…The proposed technique achieved an accuracy of 96% and demonstrated its superiority over several other techniques used in some work[20] (94% & 34% in best and worst scenario) &[21] (91.43%).…”

Bridging the gap between microarray technology and routine clinical diagnostics: a random forest approach to the gene expression profile dimensionality reduction

“…Un for tu na te ly, the num ber of mul ti va ria te fi l te rs is not ve ry lar ge. Be si des mu tual in for ma tion ba sed ap proac hes (6,7,8), cer tain sta tis ti cal and mac hi ne lear ni ng/prediction too ls cou ld be used for mul ti va ria te se lec tion of re le va nt ge nes, i.e. di men sio na li ty re duc tion.…”

“…Od oso bi tog bi znače nja bi lo is pi ti va nje prik lad nos ti od re đi va nja se rum ske kon cen tra ci je cis ta ti na C u svr hu raz li ko va nja AML i ALL. Ko nač no, ve ći na ge na nab ro je nih u Tab li ci 1 su ta ko đer pre poz na ti od stra ne dru gih au to ra (7,8,23) i ve ći na njih je ko ris ti la uni va ri jat ne fi l tre. Ova či nje ni ca uka zu je da se prob lem raz li ko va nja AML/ALL mo že sves ti na uni va ri jatni ili li near no raz dvo ji vi prob lem (Sli ke 2a i 2b).…”

“…It wou ld be es pe cial ly im por ta nt to exa mi ne the sui ta bi li ty of se rum cysta tin C con cen tra tion de ter mi na tion for AML/ALL dis cri mi na tion. Fi nal ly, mo st of the ge nes lis ted in Tab le 1 we re fou nd by ot her aut ho rs (7,8,23), and mo st of them used uni va ria te fi l te rs. This fa ct sug ges ts that AML/ALL prob lem cou ld be re du ced to a uni va ria te or li near ly se pa rab le prob lem (Figu res 2a and 2b).…”

“…Na ža lo st, broj mul ti va ri jat nih fi l tra ni je baš ve lik. Osim na uza jam noj in for ma ci ji zas no va nih pris tu pa (6,7,8), od re đeni ala ti za sta tis tič ko i stroj no uče nje/predviđanje mo gu se is ko ris ti ti za mul ti va ri jat ni oda bir re le van tnih ge na, tj. za sma nje nje di men zio nal nos ti.…”