Cros-Organism Annotation Prediction through Deep Learning Algorithms

Feroce, Marcello (2017) Cros-Organism Annotation Prediction through Deep Learning Algorithms. [Laurea], Università di Bologna, Corso di Studio in Ingegneria e scienze informatiche [L-DM270] - Cesena
Documenti full-text disponibili:
[img] Documento PDF (Thesis)
Disponibile con Licenza: Creative Commons Attribuzione - Non commerciale - Condividi allo stesso modo 3.0

Download (1MB)

Abstract

Studying how genes or proteins influence humans and other species' lives is paramount. To study that, it's necessary to know which functional properties are specific for each gene or protein. The association between one gene or protein and a functional properties is called annotation. An annotion can be 0 or 1. 1 means that gene or protein contributes to the activation of a certain functional property. Functional properties are referred by terms, which are strings that belong to ontologies. This work aim is to predict novel gene annotations for little know species such as Bos Taurus. To predict such annotations, a model, built using deep learning, is used. This model is trained using well know species as Mus Musculus or Homo Sapiens. Every predicted annotation has its own likelihood, that tells about how much the prediction is close to a 0 or a 1. Final accuracy can be evaluated fixing a certain value of likelihood, so that all the considered annotations have a likelihood greater or equal than the fixed one. The obtained accuracy is quite high but not enought to be used in a professional way, although it offers a nice cue for future research.

Abstract
Tipologia del documento
Tesi di laurea (Laurea)
Autore della tesi
Feroce, Marcello
Relatore della tesi
Correlatore della tesi
Scuola
Corso di studio
Indirizzo
Curriculum scienze e tecnologie informatiche
Ordinamento Cds
DM270
Parole chiave
Deep learning,Neural networks,Bioinformatics,Annotation prediction,gene ontology,genomic,tensorflow
Data di discussione della Tesi
5 Ottobre 2017
URI

Altri metadati

Statistica sui download

Gestione del documento: Visualizza il documento

^