In this chapter, we are interested in proteins classification starting from their primary structures. The goal is to automatically affect proteins sequences to their families. The main originality of the approach is that we directly apply the text categorization framework for the protein classification with very minor modifications. The main steps of the task are clearly identified: we must extract features from the unstructured dataset, we use the fixed length n-grams descriptors; we select and combine the most relevant one for the learning phase; and then, we select the most promising learning algorithm in order to produce accurate predictive model. We obtain essentially two main results. First, the approach is credible, giving accurate results with only 2-grams descriptors length. Second, in our context where many irrelevant descriptors are automatically generated, we must combine aggressive feature selection algorithms and low variance classifiers such as SVM (Support Vector Machine).
The Protein Classification Context
Biologists play a central role in the classification of individuals such as animals, plants, genes, proteins, etc. However, the great amount of biological data such as proteins, DNA, RNA etc. involves a strong need for the intervention of other research tools and techniques in order to help these biologists, mainly because the manual classification has become almost impossible.
Among these tasks, we are interested in the protein classification. We want to propose a framework where the classification process relies mainly on the primary description of the proteins. Proteins can be represented as variable length sequences, typically several hundred characters long, from the alphabet of 20 amino acids (Figure 1).
Key Terms in this Chapter
Protein Classification: A process which enables to build a predictive model that automatically assign the functional family of a protein sequence from its description.
Feature Selection: The aim of the feature selection is to select the relevant features for a data mining task e.g. the more efficient predictive variable for a predictive model.
Feature Construction: Feature construction is a process which builds intermediate features from the original descriptors in a dataset. The aim is to build more efficient features for a machine data mining task.
Data Mining: The application of analytical methods and tools to data for the purpose of identifying patterns and relationships such as classification, prediction, estimation, or affinity grouping.
Predictive Model: A predictive model is an equation or set of rules that makes it possible to predict an unseen value (the class attribute) from other, known values (descriptors).
Text Categorization: A process which enables to build a predictive model that automatically assign the label to a textual document from its description.f