CBRC Header
  • In search of cell-line specific predictive signatures for enhancers

    Dimitrios Kleftogiannis, Panos Kalnis and Vladimir B. Bajic

    Deciphering DNA cis-regulatory element interactions is critical for understanding distinct gene expression programs that govern cellular development, morphology and function. In particular, understanding functional mechanisms of enhancers and discovering enhancer sequence properties will shed light to the incomplete picture of understanding gene expression programs that characterize normal or pathogenic cellular conditions. In this report, we investigate the capability of identifying enhancer signatures based on state-of-the-art computational techniques. First, we assess the effectiveness of several existing approaches for feature selection and next by combining the best-performing sets we report cell-line specific predictive signatures as well as common fingerprints that characterize the broad category of enhancers. Our experimentation using data from ENCODE and FANTOM5 projects shows that satisfactory generalization capabilities can be achieved using the identified global enhancer signatures. In addition, our findings re-confirm the importance of specific dinucleotide motifs and other features that complement the existing knowledge about enhancer’s properties as obtained from experimental procedures.
  • Contents :
    Workflow: FS_framework
  • This work is part of a joint project between CBRC and InfoCloud at King Abdullah University of Science and Technology

  • Should be Empty: