4144

3915

كارشناسي ارشد

معماري كامپيوتر

اصفهان: دانشگاه صنعتي اصفهان، دانشكده برق و كامپيوتر

1386

ده، 95، [II]ص.: مصور، جدول، نمودار

ص. ع. به فارسي و انگليسي

محمد حسين سرايي، محمد داورپناه جزي

05/06/87

آقافخر ميرلوحي، رسول موسوي

1396/03/06

كتابنامه

برق و كامپيوتر

مهندسي برق و كامپيوتر

ID3915

به فارسي و انگليسي: قابل رؤيت در نسخه ديجيتال

AbstractBioinformatics is the field of science in which biology computer science andinformation technology merge to form a single discipline The ultimate goal of the field is toenable the discovery of new biological insights as well as to create a global perspective fromwhich unifying principles in biology can be established The rationale for applyingcomputational approaches to facilitate the understanding of various biological processesincludes a more global perspective in experimental design and the ability to capitalize on theemerging technology of data mining One of the main aims of current genetics research is todiscover functional relationship between genotype and phenotype Identifying the causalgenetic variants and their functional patterns may greatly facilitate the preventive and diagnosisand biochemical understanding of genetic diseases This so called gene mapping During recent years there has been growing interest in using data mining methods in genemapping motivated by the lack of success of the traditional approaches for complex diseases and also by the intriguing possibility of simultaneous detection of multiple loci The datamining methods for linkage disequilibrium mapping can be categorized into three groupsincluding classification methods clustering techniques and methods based on the discovery oftypical haplotypes Classification methods work well in modeling interaction Several of theclassification methods produce a set of interacting loci that best predict the phenotype However a straightforward application of classification methods to large numbers of markershas a potential risk picking up randomly In this thesis we present a technique to improveclassification of case and control individuals At first important marker set is selected byHapMiner method then modeling is accomplished by classification and regression tree CART method on these markers This approach has tested on real dataset Type 1 Diabetes Disease Accuracy improvement is gained by modeling on important markers in population of haploypewith respect to modeling on all markers The association gene mapping methods based on the haplotype clustering analysis are vastlyused to localize a mutation in a gene sequence In many cases the locations that are found basedon these methods have large errors In this work we present a robust technique to lower themean error of the association gene mapping in the haplotype clustering analysis In thistechnique we utilize the information gain to select a set of important features i e markers that are used in the clustering process In other words each marker is assigned a rank and thenthe high ranked markers are fed into the HapMiner algorithm for localizing the disease Inorder to justify the proposed approach We have applied the performance of our technique on aset of simulated dataset The experiments show a significant reduction in the mean error of thegene mapping

محمد حسين سرايي، محمد داورپناه جزي

آقافخر ميرلوحي، رسول موسوي