<i>RANKS</i>: a flexible tool for node label ranking and classification in biological networks

Valentini, Giorgio; Armano, Giuliano; Frasca, Marco; Lin, Jianyi; Ré, Matteo

doi:10.1093/bioinformatics/btw235

Cited by 35 publications

(28 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We included the classical diffusion raw, a weighted approach version gm and two statistically normalised scores (mc and z), as implemented in diffuStats. In the scope of positive-unlabelled learning [30,31], we included the kernelised scores knn and the linear decayed wsld from RANKS [32]. Closing this category, we implemented the bagging Support Vector Machine approach from ProDiGe1 [33], here bagsvm.…”

Section: Selection Of Methods For Investigationmentioning

confidence: 99%

Benchmarking network propagation methods for disease gene identification

Picart‐Armada

Barrett

Willé

et al. 2018

Preprint

View full text Add to dashboard Cite

BackgroundIn-silico identification of potential disease genes has become an essential aspect of drug target discovery. Recent studies suggest that one powerful way to identify successful targets is through the use of genetic and genomic information. Given a known disease gene, leveraging intermolecular connections via networks and pathways seems a natural way to identify other genes and proteins that are involved in similar biological processes, and that can therefore be analysed as additional targets.ResultsHere, we systematically tested the ability of 12 varied network-based algorithms to identify target genes and cross-validated these using gene-disease data from Open Targets on 22 common diseases. We considered two biological networks, six performance metrics and compared two types of input gene-disease association scores. We also compared several cross-validation schemes and showed that different choices had a remarkable impact on the performance estimates. When seeding biological networks with known drug targets, we found that machine learning and diffusion-based methods are able to find novel targets, showing around 2-4 true hits in the top 20 suggestions. Seeding the networks with genes associated to disease by genetics resulted in poorer performance, below 1 true hit on average. We also observed that the use of a larger network, although noisier, improved overall performance.ConclusionsWe conclude that machine learning and diffusion-based prioritisers are suited for drug discovery in practice and improve over simpler neighbour-voting methods. We also demonstrate the large effect of several factors on prediction performance, especially the validation strategy, input biological network, and definition of seed disease genes.

show abstract

Section: Selection Of Methods For Investigationmentioning

confidence: 99%

Benchmarking network propagation methods for disease gene identification

Picart‐Armada

Barrett

Willé

et al. 2018

Preprint

View full text Add to dashboard Cite

show abstract

“…For instance, RANKS [6] performs node prioritization on some label or property by using kernelized score functions, taking into account both the global structure of the network and the neighborhood of the query nodes. Other approaches, like SVD-phy , try to find functional associations between genes based on their phylogenetic distributions [7].…”

Section: Resultsmentioning

confidence: 99%

ProphTools: general prioritization tools for heterogeneous biological networks

et al. 2017

View full text Add to dashboard Cite

BackgroundNetworks have been proven effective representations for the analysis of biological data. As such, there exist multiple methods to extract knowledge from biological networks. However, these approaches usually limit their scope to a single biological entity type of interest or they lack the flexibility to analyze user-defined data.ResultsWe developed ProphTools, a flexible open-source command-line tool that performs prioritization on a heterogeneous network. ProphTools prioritization combines a Flow Propagation algorithm similar to a Random Walk with Restarts and a weighted propagation method. A flexible model for the representation of a heterogeneous network allows the user to define a prioritization problem involving an arbitrary number of entity types and their interconnections. Furthermore, ProphTools provides functionality to perform cross-validation tests, allowing users to select the best network configuration for a given problem. ProphTools core prioritization methodology has already been proven effective in gene-disease prioritization and drug repositioning. Here we make ProphTools available to the scientific community as flexible, open-source software and perform a new proof-of-concept case study on long noncoding RNAs (lncRNAs) to disease prioritization.ConclusionsProphTools is robust prioritization software that provides the flexibility not present in other state-of-the-art network analysis approaches, enabling researchers to perform prioritization tasks on any user-defined heterogeneous network. Furthermore, the application to lncRNA-disease prioritization shows that ProphTools can reach the performance levels of ad hoc prioritization tools without losing its generality.

show abstract

“…The benchmark methods adopted in [47] are the following: GBA, family of algorithms relying upon the guilt-by-association rule, which allows making predictions exploiting the interacting genes, under the assumption that interacting genes are likely to share similar functions [69,70]; RW, random walks algorithm [71]; RWR, random walks with restart which takes into account that after many steps the walker may forget the prior information coded in the initial probability vector, and accordingly the algorithm allows the walker to restart from its initial condition with a given probability θ (free parameter), or to move another random walk step with probability 1 − θ; kernelized score functions, extending the similarity to non neighboring nodes by adopting a suitable kernel matrix [72]. The score for each gene i with regard to a given disease r is defined according to a suitable distance d(i, V r ) between i and the subset V r of positive genes for r. By varying the definition of d(i, V r ) authors obtained different scoring methods, among which the top performing was S AV : d(i, V r ) is defined as the average distance between the images in the corresponding Hilbert space of i and V r .…”

Section: Resultsmentioning

confidence: 99%