Post on 28-Dec-2015
Finding the building blocks of RNA 3-D structure using graph analysis
Finding the building blocks of RNA 3-D structure using graph analysisRomain RivireAReNa 28.03.2007Interest in RNA modellingCharacterise RNA families
Improve non-coding RNA identification in genomic data
Determine the RNA players in regulatory networks
Identify potential RNA drug targets
ProjectBackground: ribonomics is curently stuck with the secondary structure paradigm, whereas we would need high throughput tertiary data
Hypothesis: Finding the fundamental building blocks of RNA structures will reduce the complexity of RNA foldingRNA 3-D Structure interactions
From 3-D Structure to Graph :MC-Annotate
Yeast tRNA-Phe crystal structure(pdb 4TNA)Solution in 3 stepsEnumerate all the motifs
Regroup by similarity
Find the building blocksEnumerating motifs : The backbone helps !
Enumerating motifs
A motif is a set of connected nucleotides together with their interactionsEnumerating motifs
344 motifs of size 4 in tRNA
Solution in 3 stepsEnumerate all the motifs
Regroup by similarity
Find the building blocksGraph Isomorphism125430 1 0 0 11 0 1 1 00 1 0 1 00 1 1 0 11 0 0 1 0125430 1 0 1 11 0 1 1 00 1 0 0 11 1 0 0 01 0 1 0 0241350 1 0 0 11 0 1 1 00 1 0 1 00 1 1 0 11 0 0 1 0Matrix representation of the graph :21==Scan through all permutations to decide
if two graphs are isomorphic !Group motifs with isomorphism????????13Canonical labelling125430 1 0 0 11 0 1 1 00 1 0 1 00 1 1 0 11 0 0 1 0125430 1 0 1 11 0 1 1 00 1 0 0 11 1 0 0 01 0 1 0 0241350 1 0 0 11 0 1 1 00 1 0 1 00 1 1 0 11 0 0 1 0Matrix representation of the graph :21==Take the minimum through all permutations of the matrix representationGroup motifs with canonical labelling18211833212115Solution in 3 stepsEnumerate all the motifs
Regroup by similarity
Find the building blocksOccurrences of a motif in the 50S rRNA of H. marismortui
The covering graph : a mapping between motifs and edges
Type of motifsEdges
77710564430Find a small set of types of motifs that covers the most edgesBuilding blocks of the 50S?(graphs of size 4)
General graphs are not compact enough
Not usable for modelling
334 Building blocks of the 50S(cycles of sizes 3 to 6)
The colored Ribosome
The colored 16S
Covering for the PDB database
Future worksMotif discoveryBiological relevance of block-function relationships?
RNA foldingPractical usability? AcknowledgmentsFranois MajorSbastien LemieuxVronique LisiKarine St-OngePhilippe ThibaultPatrick GendronMartin LaroseAll other lab members
Thank youThe ResultsWe applied the method to the large ribosomal subunit
We restrict the graphs allowed for the base to cycles
We found 334 cycles that covers 90% of the structure.Canonical labelingCanonical label of a graph : Take the minimum of the matrices over all the permutations
Property : 2 graphs have the same canonical label if and only if they are isomorphicSecond step : group motifsGroup together motifs which are identicalDone with canonical labellingIdea : associate a string to each graph such that two graphs are associated with the same string if and only if they are identical (isomorphic).Difficult problem well studiedPotentially highly computation time.The construction of a bipartite graph
Conformational space too large3^n . 10^n where n is the size of the structure
10^14 op/s world fastest computerLetter A : GC base paire
Letter Sarcin Riccin
Letter GNRA
Letter crossing
Letter GNRA (2)
Interests in RNARNA is a very important medium in the transfer of genetic information
Convey information through its structure in addition to its sequence
Example : miRNAmiRNA Pathway
Structure of an miRNA
Similar secondary structure
But, only one miRNA functional !RNA 3-D Structure ineractions3 main types of molecular interactions :Phosphodiester linkBase pairingBase stackingRNA 3-D Structure interactions : Backbone connectivity
RNA 3-D Structure interactions : Base pairing types
RNA 3-D Structure interations : Base stacking types