Beyond the 'best' match: machine learning annotation of protein sequences by integration of different sources of information

Isofunctional Protein Subfamily Detection Using Data Integration and Spectral Clustering

As increasingly more genomes are sequenced, the vast majority of proteins may only be annotated computationally, given experimental investigation is extremely costly. This highlights the need for computational methods to determine protein functions quickly and reliably. We believe dividing a protein family into subtypes which share specific functions uncommon to the whole family reduces the function annotation problem's complexity. Hence, this work's purpose is to detect isofunctional subfamilies inside a family of unknown function, while identifying differentiating residues. Similarity between protein pairs according to various properties is interpreted as functional similarity evidence. Data are integrated using genetic programming and provided to a spectral clustering algorithm, which creates clusters of similar proteins. The proposed framework was applied to well-known protein families and to a family of unknown function, then compared to ASMC. Results showed our fully automated technique obtained better clusters than ASMC for two families, besides equivalent results for other two, including one whose clusters were manually defined. Clusters produced by our framework showed great correspondence with the known subfamilies, besides being more contrasting than those produced by ASMC. Additionally, for the families whose specificity determining positions are known, such residues were among those our technique considered most important to differentiate a given group. When run with the crotonase and enolase SFLD superfamilies, the results showed great agreement with this gold-standard. Best results consistently involved multiple data types, thus confirming our hypothesis that similarities according to different knowledge domains may be used as functional similarity evidence. Our main contributions are the proposed strategy for selecting and integrating data types, along with the ability to work with noisy and incomplete data; domain knowledge usage for detecting subfamilies in a family with different specificities, thus reducing the complexity of the experimental function characterization problem; and the identification of residues responsible for specificity.

Why do Sequence Signatures Predict Enzyme Mechanism? Homology versus Chemistry

First, we identify InterPro sequence signatures representing evolutionary relatedness and, second, signatures identifying specific chemical machinery. Thus, we predict the chemical mechanisms of enzyme-catalyzed reactions from catalytic and non-catalytic subsets of InterPro signatures. We first scanned our 249 sequences using InterProScan and then used the MACiE database to identify those amino acid residues that are important for catalysis. The sequences were mutated in silico to replace these catalytic residues with glycine and then again scanned using InterProScan. Those signature matches from the original scan that disappeared on mutation were called catalytic. Mechanism was predicted using all signatures, only the 78 “catalytic” signatures, or only the 519 “non-catalytic” signatures. The non-catalytic signatures gave indistinguishable results from those for the whole feature set, with precision of 0.991 and sensitivity of 0.970. The catalytic signatures alone gave less impressive predictivity, with precision and sensitivity of 0.791 and 0.735, respectively. These results show that our successful prediction of enzyme mechanism is mostly by homology rather than by identifying catalytic machinery.

Enzyme informatics

Over the last 50 years, sequencing, structural biology and bioinformatics have completely revolutionised biomolecular science, with millions of sequences and tens of thousands of three dimensional structures becoming available. The bioinformatics of enzymes is well served by, mostly free, online databases. BRENDA describes the chemistry, substrate specificity, kinetics, preparation and biological sources of enzymes, while KEGG is valuable for understanding enzymes and metabolic pathways. EzCatDB, SFLD and MACiE are key repositories for data on the chemical mechanisms by which enzymes operate. At the current rate of genome sequencing and manual annotation, human curation will never finish the functional annotation of the ever-expanding list of known enzymes. Hence there is an increasing need for automated annotation, though it is not yet widespread for enzyme data. In contrast, functional ontologies such as the Gene Ontology already profit from automation. Despite our growing understanding of enzyme structure and dynamics, we are only beginning to be able to design novel enzymes. One can now begin to trace the functional evolution of enzymes using phylogenetics. The ability of enzymes to perform secondary functions, albeit relatively inefficiently, gives clues as to how enzyme function evolves. Substrate promiscuity in enzymes is one example of imperfect specificity in protein-ligand interactions. Similarly, most drugs bind to more than one protein target. This may sometimes result in helpful polypharmacology as a drug modulates plural targets, but also often leads to adverse side-effects. Many chemoinformatics approaches can be used to model the interactions between druglike molecules and proteins in silico. We can even use quantum chemical techniques like DFT and QM/MM to compute the structural and energetic course of enzyme catalysed chemical reaction mechanisms, including a full description of bond making and breaking.

EnzML: multi-label prediction of enzyme classes using InterPro signatures

Background

Manual annotation of enzymatic functions cannot keep up with automatic genome sequencing. In this work we explore the capacity of InterPro sequence signatures to automatically predict enzymatic function.

Results

We present EnzML, a multi-label classification method that can efficiently account also for proteins with multiple enzymatic functions: 50,000 in UniProt. EnzML was evaluated using a standard set of 300,747 proteins for which the manually curated Swiss-Prot and KEGG databases have agreeing Enzyme Commission (EC) annotations. EnzML achieved more than 98% subset accuracy (exact match of all correct Enzyme Commission classes of a protein) for the entire dataset and between 87 and 97% subset accuracy in reannotating eight entire proteomes: human, mouse, rat, mouse-ear cress, fruit fly, the S. pombe yeast, the E. coli bacterium and the M. jannaschii archaebacterium. To understand the role played by the dataset size, we compared the cross-evaluation results of smaller datasets, either constructed at random or from specific taxonomic domains such as archaea, bacteria, fungi, invertebrates, plants and vertebrates. The results were confirmed even when the redundancy in the dataset was reduced using UniRef100, UniRef90 or UniRef50 clusters.

Conclusions

InterPro signatures are a compact and powerful attribute space for the prediction of enzymatic function. This representation makes multi-label machine learning feasible in reasonable time (30 minutes to train on 300,747 instances with 10,852 attributes and 2,201 class values) using the Mulan Binary Relevance Nearest Neighbours algorithm implementation (BR-kNN).

Towards a semi-automatic functional annotation tool based on decision-tree techniques

Background

Due to the continuous improvements of high throughput technologies and experimental procedures, the number of sequenced genomes is increasing exponentially. Ultimately, the task of annotating these data relies on the expertise of biologists. The necessity for annotation to be supervised by human experts is the rate limiting step of the data analysis. To face the deluge of new genomic data, the need for automating, as much as possible, the annotation process becomes critical.

Results

We consider annotation of a protein with terms of the functional hierarchy that has been used to annotate Bacillus subtilis and propose a set of rules that predict classes in terms of elements of the functional hierarchy, i.e., a class is a node or a leaf of the hierarchy tree. The rules are obtained through two decision-trees techniques: first-order decision-trees and multilabel attribute-value decision-trees, by using as training data the proteins from two lactic bacteria: Lactobacillus sakei and Lactobacillus bulgaricus. We tested the two methods, first independently, then in a combined approach, and evaluated the obtained results using hierarchical evaluation measures. Results obtained for the two approaches on both genomes are comparable and show a good precision together with a high prediction rate. Using combined approaches increases the recall and the prediction rate.

Conclusion

The combination of the two approaches is very encouraging and we will further refine these combinations in order to get rules even more useful for the annotators. This first study is a crucial step towards designing a semi-automatic functional annotation tool.