Neki statistički aspekti prepoznavanja motiva

master's thesis

• Author: Ajka Relja
• Mentor: Pavle Goldstein (mentor)
• Committee member: Pavle Goldstein (predsjednik povjerenstva)
• Committee member: Filip Najman (član povjerenstva)
• Committee member: Franka Miriam Bruckler (član povjerenstva)
• Committee member: Bojan Basrak (član povjerenstva)
• Granter: University of Zagreb; Faculty of Science; (Department of Mathematics); Zagreb
• Defense date and country: 2014-07-14, Croatia
• Scientific / art field, discipline and subdiscipline: NATURAL SCIENCES; Mathematics

Abstract

U ovom diplomskom radu analiziramo distribuciju maksimalnih scorova dobivenih za stvarni enzim iz familije GDSL hidrolaza na biljci A. thaliani. Tema nas je zaintrigirala zbog činjenice što takva distribucija ne odgovara Gumbelovoj distribuciji iz familije distribucija ekstremnih vrijednosti koja je teorijski opravdana. Izračunavamo scorove za upit koristeći se metodom klizećeg prozora i specifičnom matricom težina i analiziramo njihovu distribuciju. Ono što primjećujemo je da uz određenu korekciju scorova s obzirom na duljinu niza nad kojim su izračunati, uspijevamo dobiti Gumbelovu razdiobu. Na kraju provjeravamo kako je korekcija utjecala na scorove pozitivaca, odnosno onih enzima za koje znamo da su u funkcionalnoj vezi s našim enzimom čije smo varijante željeli pronaći u proteomu biljke. To je od iznimne važnosti prvenstveno zbog činjenice što želimo da korekcijom oni enzimi koji su u vezi sa upitom i dalje imaju veći score od onih koji nisu. Pokaže se da naša korekcija ne kvari takav odnos.

Abstract (english)

In this thesis we have analysed the distribution of the maximum scores obtained for an actual enzyme from GDSL family of hydrolases on the plant A. thaliana. This is an intriguing topic, since such distribution does not match Gumbel distribution, as it should, according to relevant theoretical results. For the purpose of analysing such distribution, we have calculated scores for query using the sliding window protocol and PSSM matrix. What we have noticed is that a certain correction of scores with respect to the length of the protein, gives Gumbel distribution. Finally we check how the correction affected scores of positive matches, i.e. those enzymes that are in functional relation with our query whose variants we wanted to find in the proteome. That is rather important, primarily since we want enzymes which are related to the query to keep higher scores than those that are not. We show that our correction keeps such a relationship.

• Keywords: analiza distribucije maksimalnih scorova; Gumbelova razdioba; enzim iz familije GDSL hidrolaza na biljci A. thaliani
• Keywords (english): the distribution of the maximum scores analysis; Gumbel distribution; enzyme from GDSL family of hydrolases on the plant A. thaliana
• Language: croatian
• URN:NBN: urn:nbn:hr:217:401665
• Study programme: Title: Mathematical Statistics; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra matematike (magistar/magistra matematike)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Open access
• Created on: 2019-02-05 13:23:48