Filtern
Volltext vorhanden
- ja (3)
Gehört zur Bibliographie
- ja (3)
Dokumenttyp
Sprache
- Englisch (3)
Schlagworte
- Disease gene prioritization (1)
- Protein function prediction (1)
- RNA-binding proteins (1)
- bacterial secretion (1)
- characterization (1)
- identification (1)
- pathogens (1)
Institut
An expanded evaluation of protein function prediction methods shows an improvement in accuracy
(2016)
Background
A major bottleneck in our understanding of the molecular underpinnings of life is the assignment of function to proteins. While molecular experiments provide the most reliable annotation of proteins, their relatively low throughput and restricted purview have led to an increasing role for computational function prediction. However, assessing methods for protein function prediction and tracking progress in the field remain challenging.
Results
We conducted the second critical assessment of functional annotation (CAFA), a timed challenge to assess computational methods that automatically assign protein function. We evaluated 126 methods from 56 research groups for their ability to predict biological functions using Gene Ontology and gene-disease associations using Human Phenotype Ontology on a set of 3681 proteins from 18 species. CAFA2 featured expanded analysis compared with CAFA1, with regards to data set size, variety, and assessment metrics. To review progress in the field, the analysis compared the best methods from CAFA1 to those of CAFA2.
Conclusions
The top-performing methods in CAFA2 outperformed those from CAFA1. This increased accuracy can be attributed to a combination of the growing number of experimental annotations and improved methods for function prediction. The assessment also revealed that the definition of top-performing algorithms is ontology specific, that different performance metrics can be used to probe the nature of accurate predictions, and the relative diversity of predictions in the biological process and human phenotype ontologies. While there was methodological improvement between CAFA1 and CAFA2, the interpretation of results and usefulness of individual methods remain context-dependent.
Many pathogenic bacteria utilize specialized secretion systems to deliver proteins called effectors into eukaryotic cells for manipulation of host pathways. The vast majority of known effector targets are host proteins, whereas a potential targeting of host nucleic acids remains little explored. There is only one family of effectors known to target DNA directly, and effectors binding host RNA are unknown. Here, we take a two-pronged approach to search for RNA-binding effectors, combining biocomputational prediction of RNA-binding domains (RBDs) in a newly assembled comprehensive dataset of bacterial secreted proteins, and experimental screening for RNA binding in mammalian cells. Only a small subset of effectors were predicted to carry an RBD, indicating that if RNA targeting was common, it would likely involve new types of RBDs. Our experimental evaluation of effectors with predicted RBDs further argues for a general paucity of RNA binding activities amongst bacterial effectors. We obtained evidence that PipB2 and Lpg2844, effector proteins of Salmonella and Legionella species, respectively, may harbor novel biochemical activities. Our study presenting the first systematic evaluation of the RNA-targeting potential of bacterial effectors offers a basis for discussion of whether or not host RNA is a prominent target of secreted bacterial proteins.
RNA-binding proteins (RBPs) have been established as core components of several post-transcriptional gene regulation mechanisms. Experimental techniques such as cross-linking and co-immunoprecipitation have enabled the identification of RBPs, RNA-binding domains (RBDs) and their regulatory roles in the eukaryotic species such as human and yeast in large-scale. In contrast, our knowledge of the number and potential diversity of RBPs in bacteria is poorer due to the technical challenges associated with the existing global screening approaches. We introduce APRICOT, a computational pipeline for the sequence-based identification and characterization of proteins using RBDs known from experimental studies. The pipeline identifies functional motifs in protein sequences using position-specific scoring matrices and Hidden Markov Models of the functional domains and statistically scores them based on a series of sequence-based features. Subsequently, APRICOT identifies putative RBPs and characterizes them by several biological properties. Here we demonstrate the application and adaptability of the pipeline on large-scale protein sets, including the bacterial proteome of Escherichia coli. APRICOT showed better performance on various datasets compared to other existing tools for the sequence-based prediction of RBPs by achieving an average sensitivity and specificity of 0.90 and 0.91 respectively. The command-line tool and its documentation are available at https://pypi.python.org/pypi/bio-apricot.