Identification of Functionally Related Enzymes by Learning-to-Rank Methods

Abstract

Enzyme sequences and structures are routinely used in the biological sciences as queries to search for functionally related enzymes in online databases. To this end, one usually departs from some notion of similarity, comparing two enzymes by looking for correspondences in their sequences, structures or surfaces. For a given query, the search operation results in a ranking of the enzymes in the database, from very similar to dissimilar enzymes, while information about the biological function of annotated database enzymes is ignored. In this work, we show that rankings of that kind can be substantially improved by applying kernel-based learning algorithms. This approach enables the detection of statistical dependencies between similarities of the active cleft and the biological function of annotated enzymes. This is in contrast to search-based approaches, which do not take annotated training data into account. Similarity measures based on the active cleft are known to outperform sequence-based or structure-based measures under certain conditions. We consider the Enzyme Commission (EC) classification hierarchy for obtaining annotated enzymes during the training phase. The results of a set of sizeable experiments indicate a consistent and significant improvement for a set of similarity measures that exploit information about small cavities in the surface of enzymes.

Multi-target prediction: a unifying view on problems and methods

“…However, the multi-task learning terminology is rarely used in the dyadic prediction literature. Dyadic prediction problems emerge in a variety of application domains, including product recommendation, social network analysis, drug design, various bioinformatics applications, and game playing (Basilico and Hofmann, 2004;Park and Chu, 2009;Stock et al, 2014;Ben-Hur and Noble, 2005;Kashima et al, 2009;Pelossof et al, 2015;Pahikkala et al, 2010Pahikkala et al, , 2013.…”

Infinite-Label Learning with Semantic Output Codes

“…Learning to rank [28,29] and the cold-start problem in recommendation system [30,31,32] explicitly attempt to tackle unseen labels (e.g., queries and items, respectively) thanks to the label descriptions that are often vectorial. Zero-shot learning shares the same spirit [1,2,5,3,4,7,6,8,33,34] and we particularly study the infinite-label learning in this paper.…”

Exact and efficient top-K inference for multi-target prediction by querying separable linear relational models

“…Models of this type are frequently encountered in several subfields of machine learning, such as pairwise learning, content-based filtering, biological network inference, dyadic prediction and conditional ranking. They are particularly popular in certain domains of bioinformatics, such as the prediction of protein-protein interactions [6,26,48], enzyme function prediction [43] and proteochemometrics [24,28,53].…”