dc.contributor.advisor | Sætrom, Pål | |
dc.contributor.author | Heggedal, Håkon | |
dc.date.accessioned | 2019-09-11T10:56:00Z | |
dc.date.created | 2015-07-01 | |
dc.date.issued | 2015 | |
dc.identifier | ntnudaim:10665 | |
dc.identifier.uri | http://hdl.handle.net/11250/2615818 | |
dc.description.abstract | Micro RNAs (miRNAs) are a group of apx. 22 nucleotide (nt) non-coding RNA sequences, playing an important role in gene regulation. The set of known miRNA for humans are grouped as high confidence (HC) miRNA, and low confidence (LC) miRNA. The HC miRNA are considered good, but the set of LC miRNA are may have several false positives.
My SVM classifier for classifying LC miRNA shows separate between good and bad LC miRNA. It finds several LC miRNAs which are likely to be true miRNA, and several other LC miRNA which are very likely not to be miRNA. I have also classified candidate miRNA, which does not get better results than the well known mirDeep2.
I have also made two new features that are used in the classification, which both separate High confidence and non-hairpin structures, but I cannot prove that using them gives any better results. | en |
dc.language | eng | |
dc.publisher | NTNU | |
dc.subject | Informatikk, Intelligente systemer | en |
dc.title | Classifying low confidence miRNA as true and false miRNA | en |
dc.type | Master thesis | en |
dc.source.pagenumber | 79 | |
dc.contributor.department | Norges teknisk-naturvitenskapelige universitet, Fakultet for informasjonsteknologi og elektroteknikk,Institutt for datateknologi og informatikk | nb_NO |
dc.date.embargoenddate | 10000-01-01 | |