Information-Theoretic Evaluation for Computational Biomedical Ontologies (eBook)

Wyatt Travis Clark (Autor)

eBook Download: PDF

2014 | 2014
VII, 46 Seiten
Springer International Publishing (Verlag)
978-3-319-04138-4 (ISBN)

Infos

The development of effective methods for the prediction of ontological annotations is an important goal in computational biology, yet evaluating their performance is difficult due to problems caused by the structure of biomedical ontologies and incomplete annotations of genes. This work proposes an information-theoretic framework to evaluate the performance of computational protein function prediction. A Bayesian network is used, structured according to the underlying ontology, to model the prior probability of a protein's function. The concepts of misinformation and remaining uncertainty are then defined, that can be seen as analogs of precision and recall. Finally, semantic distance is proposed as a single statistic for ranking classification models. The approach is evaluated by analyzing three protein function predictors of gene ontology terms. The work addresses several weaknesses of current metrics, and provides valuable insights into the performance of protein function prediction tools.

Inhalt

Preface 5
Contents 6
1 Introduction 7
1.1 Background 10
1.2 Protein Function Prediction Scenarios 12
1.3 State of the Art Methods 13
References 13
2 Methods 18
2.1 Calculating the Joint Probability of a Graph 18
2.1.1 Calculating the Information Content of a Graph 21
2.1.2 Comparing Two Annotation Graphs 22
2.1.3 Measuring the Quality of Function Prediction 23
2.1.4 Weighted Metrics 25
2.1.5 Semantic Distance 25
2.1.6 Precision and Recall 26
2.1.7 Supplementary Evaluation Metrics 27
2.1.8 Additional Topological Metrics 30
2.2 Confusion Matrix Interpretation of ru and mi 30
2.3 Annotation Models 31
2.3.1 The Naïve Model 31
2.3.2 The BLAST Model 32
2.3.3 The GOtcha Model 32
References 32
3 Experiments and Results 34
3.1 Average Information Content of a Protein 34
3.2 Comparative Examples of Calculating Information Content 35
3.3 Two-Dimensional Plots 38
3.4 Comparisons of Single Statistics 40
References 45
4 Discussion 47
References 48
Index 49

Produkt-Details

• Erscheint lt. Verlag: 9.1.2014
• Reihe/Serie: SpringerBriefs in Computer Science
• Reihe/Serie: SpringerBriefs in Computer Science
• Zusatzinfo: VII, 46 p. 12 illus., 6 illus. in color.
• Verlagsort: Cham
• Sprache: englisch
• Themenwelt: Mathematik / Informatik ► Informatik ► Programmiersprachen / -werkzeuge
• Themenwelt: Studium ► 2. Studienabschnitt (Klinik) ► Humangenetik
• Themenwelt: Naturwissenschaften ► Biologie
• Schlagworte: Algorithm analysis and problem complexity • algorithm development • Information Theory • Ontology • Protein function prediction • Semantic Evaluation
• ISBN-10: 3-319-04138-X / 331904138X
• ISBN-13: 978-3-319-04138-4 / 9783319041384