fragAnchor | GPI-DB

FragAnchor: GPI-Anchored Protein Prediction Tandem System (NN+HMM)

FragAnchor is based on the tandem use of a Neural Network predictor and a Hidden Markov Model predictor. The Neural Network is used to select the potential GPI-anchored sequences and the Hidden Markov Model classifies the selected sequences according to four different levels of precision (highly probable, probable, weakly probable, potential false positive). The Hidden Markov Model proposes also up to three possible locations for the anchor/cleavage site.

fragAnchor Put your fasta sequence(s) here [Input format]:

Upload seqence file:

References

Poisson G, Chauve C, Chen X, Bergeron A. FragAnchor a large scale all Eukaryota predictor of Glycosylphosphatidylinositol-anchor in protein sequences by qualitative scoring. Genomics, Proteomics & Bioinformatics. 2007 May;5(2):121-30.

Paper using fragAnchor

Nakayasu ES, Yashunsky DV, Nohara LL, Torrecilhas AC, Nikolaev AV, Almeida IC. GPIomics: global analysis of glycosylphosphatidylinositol-anchored molecules of Trypanosoma cruzi. Mol Syst Biol. 2009;5:261. Epub 2009 Apr 7

Supporting Online Material from the paper

  • Training and Test data
  • Intergr8 proteomes analysis in GPI-DB

    Preliminary versions were presented in two posters:

    1. Hybrid system NN/HMM for large-scale GPI-anchored protein prediction. ISMB/ECCB 2004, poster I-87.
    2. Prediction of post translational GPI-anchor modification of proteins by machine-learning. HUPO 2nd Annual Congress, (abstract in Molecular & Cellular Proteomics 2(9):826, 2003).
    Questions and comments: guylaine@hawaii.edu