Skip to content

Latest commit

 

History

History
101 lines (60 loc) · 7.84 KB

index-graphclust.md

File metadata and controls

101 lines (60 loc) · 7.84 KB
layout
subsite-galaxy

Welcome to GraphClust2 Galaxy server for clustering of RNAs according to sequence and secondary structures similarities.

GraphClust2 is a workflow for scalable clustering of RNAs based on sequence and secondary structures feature. GraphClust2 is implemented within the Galaxy framework and consists a set of integrated Galaxy tools and flavors of the linear-time clustering workflow.

  1. TOC {:toc}

Get started

Interactive tours

GraphClust2 rapid start

Interactive tours are available for Galaxy and GraphClust2. To run the tours please on top panel go to Help→Interactive Tours and click on one of the tours prefixed GraphClust workflow (direct link to the basic tour). Please use your personal user-password for logging in. You can check the other tours for a more general introduction to the Galaxy interface.

Galaxy interface

Are you new to Galaxy, or returning after a long time, and looking for help to get started? Take a guided tour{:target="_top"} through Galaxy's user interface.

GraphClust2 repository

Please also refer to the GraphClust2 repository for other deployment options and manuals.

Video tutorial

This video tutorial can be helpful to get a visually comprehensive introduction on setting-up and running GraphClust2. The video starts with setting up the docker Galaxy server that can be skipped through using this server.

IMAGE ALT TEXT HERE{:target="_top"}

Workflows on this server

Below workflows can be directly accessed on the public server:

Workflow flavors

The pre-configured flavors of GraphClust2 are provided and described inside the workflows directory

Import or upload a workflow

To import or upload additional workflow flavors (e.g. from workflows directory), on the top panel go to Workflow menu. On top right side of the screen click on "Upload or import workflow" button. You can either upload workflow from your local system or by providing the URL of the workflow. Log in is necessary to access into the workflow menu. The docker galaxy instance has a pre-configured easy! info that can be found by following the interactive tour. You can download workflows from the following links

GraphClust pipeline overview

The pipeline for clustering RNA sequences and structured motif discovery is a multi-step pipeline. Overall it consists of three major phases: a) sequence based pre-clustering b) encoding predicted RNA structures as graph features c) iterative fast candidate clustering then refinement

GraphClust-2 workflow overview

Below is a coarse-grained correspondence list of GraphClust2 tool names with each step:

Stage Galaxy Tool Name Description
1 Preprocessing{:target="_top"} Input preprocessing (fragmentation)
2 fasta_to_gspan{:target="_top"} Generation of structures via RNAshapes and conversion into graphs
3 NSPDK_sparseVect{:target="_top"} Generation of graph features via NSPDK
4 NSPDK_candidateClusters{:target="_top"} min-hash based clustering of all feature vectors, output top dense candidate clusters
5 PGMA_locarna{:target="_top"},locarna{:target="_top"}, CMfinder{:target="_top"} Locarna based clustering of each candidate cluster, all-vs-all pairwise alignments, create multiple alignments along guide tree, select best subtree, and refine alignment.
6 Build covariance models{:target="_top"} create candidate model
7 Search covariance models{:target="_top"} Scan full input sequences with Infernal's cmsearch to find missing cluster members
8,9 Report results and conservation evaluations Collect final clusters and create example alignments of top cluster members
{: .table.table-striped}

Input

The input to the workflow is a set of putative RNA sequences in FASTA format. Inside the data directory within the repository, you can find examples of the input format.

Output

The output contains the predicted clusters, where similar putative input RNA sequences form a cluster. Additionally overall status of the clusters and the matching of cluster elements is reported for each cluster.

Configuring the workflows:

Please proceed with the interactive tour named GraphClust workflow step by step, available under Help->Interactive Tours Please refer to the in-wrapper help descriptions the tools documentations and the repository's FAQs for checking the important parameters.

Support & Bug Reports

You can file an github issue or find our contact information in the Backofen lab page.

References

The manuscript is currently under prepration/revision. If you find this resource useful, please cite the zenodo DOI of the repo or contact us.

  • M. Miladi, E. Sokhoyan, T. Houwaart, S. Heyne, F. Costa, R. Backofen and B. Gruening; Empowering the annotation and discovery of structured RNAs with scalable and accessible integrative clustering (under preparation/revision)
  • Milad Miladi, Björn Grüning, & Eteri Sokhoyan. BackofenLab/GraphClust-2: Zenodo. http://doi.org/10.5281/zenodo.1135094
  • The GraphClust-2 alignment-free approach is based on GraphClust methodology (S. Heyne, F. Costa, D. Rose, R. Backofen; GraphClust: alignment-free structural clustering of local RNA secondary structures; Bioinformatics, 2012) available at http://www.bioinf.uni-freiburg.de/Software/GraphClust/