Geeniekspressiooni andmete analüüsi meetodi Barcode kirjeldus ja rakendamine
Date
2016
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Käesoleva bakalaureusetöö peamised eesmärgid on üle kontrollida, kas\n\rgeeniekspressiooni andmete analüüsi meetod Barcode täiustab meetodit fRMA ja tuua\n\rerinevused visuaalselt välja.\n\rEsimene, kirjeldav osa keskendub geeniekspressiooni andmete analüüsi meetodil Barcode.\n\rBarcode'i kirjelduse käigus antakse ülevaade erinevatest Barcode'i versioonidest. Iga\n\rversiooni juures on kirjeldatud funktsionaalsused ja nende kasutamine. Põhirõhk on seejuures\n\rpandud uutele funktsionaalsustele võrreldes varasemate versioonidega.\n\rTeises, praktilises osas võrreldakse meetodeid Barcode ja fRMA (fRMA meetodi väljund\n\ron Barcode analüüsi alguspunkt). Nende kahe meetodi võrdlemiseks kasutatakse\n\rinimese geeniekspressiooni andmehulka DNA kiibi eksperimentidest. Andmehulk tähisega\n\rE-TABM-145 sisaldab 158 inimese koenäidise ekspressiooniandmeid. Kõigepealt jaotatakse\n\rneed koenäidised manuaalselt gruppidesse. Need manuaalselt loodud grupid on\n\raluseks mõlema meetodi töö hindamisele. Seejärel töödeldakse algseid andmeid nii meetodiga\n\rBarcode kui ka meetodiga fRMA. Mõlema meetodi tulemuste visualiseerimiseks\n\rja võrdlemiseks kasutatakse eraldi kahte statistilist meetodit: peakomponentanalüüs (principal\n\rcomponent analysis) ja hierarhiline klasterdamine. Mõlema statistilise meetodi\n\rväljunditele on tehtud analüüs ja võrdlus Barcode'i ja fRMA vahel. Vastavate statistiliste\n\rmeetodite väljundite võrdlusest saab järeldada, et Barcode on tõepoolest täiendab\n\rfRMA-d. Barcode võimaldab koenäidiseid apremini õigetesse klastritesse klassifitseerida -\n\rnäidised, mis tulevad samast koest on kasutades Barcode'i paremini ülejäänud näidistest\n\reraldatud kui fRMA puhul.
The main goals of this thesis is to assert whether gene expression data analysis method\n\rBarcode offers improvement over the method fRMA and to visualise the difference clearly.\n\rFirst, descriptive part of this thesis focuses on the gene expression data analysis\n\rmethod Barcode. Barcode is explained by presenting an overview of different Barcode\n\rversions. For each version a description of functionalities and possible uses are given with\n\remphasis on new functionalities, compared to the older versions.\n\rSecond, practical part of this thesis compares Barcode and fRMA method(fRMA\n\rmethod output is the starting point for Barcode analysis). To compare these two methods\n\rhuman gene expression dataset of DNA microarray experiment results is used. The\n\rdataset E-TAB-145 contains expression data from 158 human tissue samples. Tissue samples\n\rare first manually clustered to use as reference in comparison of these two methods.\n\rData is then analysed with both Barcode and fRMA. To visualise and compare the result\n\rtwo statistical methods are separately used: Principal component analysis and Hierarchical\n\rclustering. For the results of both statisical analysis methods a detailed analysis\n\ris given. In the analysis it is concluded that Barcode really does offer an improvement\n\rover fRMA. Barcode allows samples to be classified better into clusters - samples of the\n\rsame tissue type are separated better from other samples compared to fRMA.
The main goals of this thesis is to assert whether gene expression data analysis method\n\rBarcode offers improvement over the method fRMA and to visualise the difference clearly.\n\rFirst, descriptive part of this thesis focuses on the gene expression data analysis\n\rmethod Barcode. Barcode is explained by presenting an overview of different Barcode\n\rversions. For each version a description of functionalities and possible uses are given with\n\remphasis on new functionalities, compared to the older versions.\n\rSecond, practical part of this thesis compares Barcode and fRMA method(fRMA\n\rmethod output is the starting point for Barcode analysis). To compare these two methods\n\rhuman gene expression dataset of DNA microarray experiment results is used. The\n\rdataset E-TAB-145 contains expression data from 158 human tissue samples. Tissue samples\n\rare first manually clustered to use as reference in comparison of these two methods.\n\rData is then analysed with both Barcode and fRMA. To visualise and compare the result\n\rtwo statistical methods are separately used: Principal component analysis and Hierarchical\n\rclustering. For the results of both statisical analysis methods a detailed analysis\n\ris given. In the analysis it is concluded that Barcode really does offer an improvement\n\rover fRMA. Barcode allows samples to be classified better into clusters - samples of the\n\rsame tissue type are separated better from other samples compared to fRMA.