Topological and semantic Web based method for analyzing TGF-β signaling pathways

(1)

HAL Id: hal-01242893

https://hal.archives-ouvertes.fr/hal-01242893

Submitted on 14 Dec 2015

HAL is a multi-disciplinary open access

archive for the deposit and dissemination of sci-entific research documents, whether they are pub-lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.

Topological and semantic Web based method for analyzing TGF-β signaling pathways

Jean Coquet, Geoffroy Andrieux, Jacques Nicolas, Olivier Dameron, Nathalie Théret

To cite this version:

Jean Coquet, Geoffroy Andrieux, Jacques Nicolas, Olivier Dameron, Nathalie Théret. Topological and semantic Web based method for analyzing TGF-β signaling pathways. JOBIM 2015, Dec 2015, Clermont-Ferrand, France. JOBIM : Journées Ouvertes en Biologie, Informatique & Mathématiques, 2015. �hal-01242893�

(2)

RESULTS

ANALYSIS OF TGF-β SIGNALIZATION PATHWAYS THANKS

TO TOPOLOGICAL AND SEMANTIC WEB METHOD

CONTEXT

How to explore the combinatorial complexity of cell signaling?

1_{Université de Rennes 1 - IRISA/INRIA, UMR6074, 263 avenue du Général Leclerc, 35042 Rennes, Cedex, France} 2_{Université de Rennes 1 - IRSET EA 4427 SeRAIC, IFR140, 2 avenue Pr Léon Bernard, Rennes, Cedex, France}

3_{German Cancer Research Center (DKFZ), Im Neuenheimer Feld 280, 69120 Heidelberg, Germany}

Jean COQUET

1,2

, Geoﬀroy ANDRIEUX

3

, Jacques NICOLAS

1

, Olivier DAMERON

1

and Nathalie THERET

2

I

RAW DATA

ˆ TGF-β signalizations inﬂuence some genes

ˆ Other stimuli in addition to TGF-β

TGF-β _{Other stimuli} g₁ g₂ g₃ g₄ g₅ p₁ p2 p₅ p₇ p₁₁ p₁₆ p₁₅ p₁₇ p₁₂ p₈ p₆ p₃ _p 4 p₉ p₁₀ p₁₄ p₁₃ p₁₈

II

GENES INFLUENCED BY PROTEIN SETS

ˆ Signaling pathway = proteins set

ˆ Proteins set inﬂuence at least one gene

s₁ = {p₁,p₇,p₁₁,p₁₅} s₂ = {p₁,p₁₅} s₃ = {p₂,p₅,p₁₅} s₄ = {p₂,p₅,p₁₆} s₅ = {p₃,p₆,p₈,p₁₂,p₁₇} s₆ = {p₃,p₆,p₈,p₉,p₁₃} s₇ = {p₃,p₆,p₈,p₉,p₁₀,p₁₈} s₈ = {p₄,p₁₄} g₁ g₂ g₃ g₄ g₅

III

(1) {g₁,g₂} x {s₁,s₂,s₃} (2) {g₂} x {s₁,s₂,s₃,s₄} (3) {g₃,g₄} x {s₅} (4) {g₄} x {s₅,s₆,s₇} (5) {g₄,g₅} x {s₇} (6) {g₅} x {s₇,s₈} Formal Concepts

CLUSTERS OF SIMILAR GENES AND PROTEIN SETS

ˆ Identify clusters of similar genes and

protein sets with formal concept analysis

ˆ Binary matrix

ˆ Formal concepts = maximal bi-cliques

g₁ g₂ g₃ g₄ g₅ s₁ s₂ s₃ s₄ s₅ s₆ s₇ s₈ 1 1 1 0 0 0 0 0 1 1 1 1 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 1 1 1 0 0 0 0 0 0 0 1 1 Protein sets Genes

IV

ASSESS CLUSTERS' BIOLOGICAL RELEVANCE WITH SEMANTIC

HOMOGENEITY

ˆ Formal concept homogeneity = zScore of

similarity gene group

ˆ Semantic similarity based on Wang measure

[Wang et al, 2007] 0.8 0.8 0.8 0.8 0.8 0.6 0.43 0.24 0.21 Gene Ontology GO terms of g1 GO terms of g2

• zScore > 1.96 : strong similarity • zScore < -1.96 : low similarity

Semantic Similarity X

zScore(fc_i) = sim(fc_i) - m σ

zScore for each genes group:

BIOLOGICAL QUESTIONS:

Which pathways are cancer-speciﬁc?

Which mechanism is involved in the healthy → pathological transition? ˆ TGF-β = Transforming Growth Factor β

ˆ necessary in healthy situation

ˆ deleterious role in ﬁbrosis and cancer

ˆ Pleiotropic eﬀects of TGF-β are linked to the complex nature of its activation and signaling networks

ˆ Model of TGF-β signal propagation based on guarded transitions [Andrieux et al, 2014] represented in CADBIOM.

ˆ Pathway Interaction Database (PID) transform into a single uniﬁed model of signal transduction

ˆ Some formal concepts are highly connected

ˆ Gene clusters can have:

ˆ High homogeneity = genes similar inﬂuenced by same pathways

ˆ Low homogeneity = genes dissimilar inﬂuenced by same pathways

ˆ Chains of gene clusters included into one-another indicates that some inﬂuence sets are speciﬁc to some gene subsets

ˆ Validate the method on small synthetic models

ˆ Improve the method to ﬁnd better gene clusters

ˆ Integrate the overlap between proteins sets

ˆ Try diﬀerent methods of clustering

ˆ Compute the semantic homogeneity of protein sets

ˆ Take into account the structure underlying gene sets

ˆ Compute the stability of formal concepts

CONCLUSION

New method to analyze pathways: 1) Topological analysis

2) Clustering

3) Biological relevance

Model of TGF-β signal propagation

Gene clusters inﬂuenced by same pathways

with a biological relevance

References:

[Andrieux et al, 2014] Andrieux, G., Le Borgne, M., & Théret, N. (2014). An

integrative modeling framework reveals plasticity of TGF-beta signaling. BMC systems biology, 8(1), 30.

[Wang et al, 2007] Wang, J. Z., Du, Z., Payattakool, R., Philip, S. Y., & Chen, C. F. (2007). A new method to measure the semantic similarity of GO terms.

Bioinformatics, 23(10), 1274-1281.

SIGNALIZATION MODEL

ANALYSIS OF SIGNALING PATHWAYS : STEPS METHOD

IN PROGRESS

DISCUSSION

DATA ANALYSIS PROBLEM:

16,000 chains of reactions linking TGF-β to at least one of 159 target genes.

Concept lattice. Nodes are formal concepts, two nodes are linked if one is included into the other.

1 13 1 1 5 1 3 3 4 1 3 2 17 1 2 1 12 11 10 9 9 6 8 5 15 14 3 12 11 5 4 2 2 1 2 1 2 1 2 1 2 1 7 4 6 3 4 3 22 6 1 5 4 13 1 12 1 2 11 1 10 1 7 6 6 5 3 18 2 2 2 0 Homogeneity zScore -1.8 2.2 Number of genes 22 1 22