About RIDDLE   |   RIDDLE Search

 a network-based method for characterizing gene sets. It asks if an input query set is significantly "close" to a known pathway or disease set in the human functional network. RIDDLE determines functional closeness between sets by integrating three major pieces of information:

Classic Enrichment - Do the query genes and the pathway overlap significantly more than expected by random chance? Most existing methods are based solely on overlap information.

Reflective Diffusion (RD) - Using a diffusion algorithm related to a popular web search algorithm, we can rank how close every gene in the network is to a given gene set. Then we can measure if a second set is highly ranked using a performance measure, such as Area Under the Curve or Average Precision.

Local Extension (LE) - Genes that are close neighbors in the network are likely to share the same function. Thus, we can extend a gene set to include the most strongly connected neighbors. With the new extended set, we check for improved enrichment.

RD and LE are performed first on the query set, then repeated with known pathway sets. All the information is combined through a machine learning approach to determine a final Riddle Association Score (RAS) and False Discovery Rate (FDR) to assess the significance of the matched gene sets. We have found that using a quality, large-scale network greatly improves upon existing approaches for finding functional associations of gene sets. Notably, RIDDLE can find relevant functions of gene sets that are poorly or not at all annotated. A detailed description of the method is available in the paper referenced below.

 in collaboration by the Marcotte Lab at Univerisy of Texas at Austin, USA and the Lee Lab at Yonsei University, Korea.

  please go to the search page and enter list of human genes, in the box provided, using Entrez gene IDs (e.g. 7157 for TP53).

Current pathway statistics :
DB# pathways
GO Biological Process (GOBP)2598

Reference :
RIDDLE: Reflective diffusion and local extension reveal functional associations for unannotated gene sets via proximity in a gene network, Peggy I Wang, Sohyun Hwang, Rodney P. Kincaid, Christopher S. Sullivan, Insuk Lee, Edward M. Marcotte Genome Biology 13:R125 (2012)

Back to Main