Click for larger image

Learning, Observing and Outputting Protein Patterns (LOOPP)

The previous version of LOOPP (2008) is still hosted at as a part of BioHPC suite.

The new version of LOOPP (2010) can be accessed at  

The work to add the new version of LOOPP to the BioHPC suite is underway and hopefully the new version will be available on BioHPC Cornell server soon.

Additionally, LOOPP is fully open-source software.  It is MPI-based software written to be platform agnostic, but is primarily tested on Linux-based clusters.  Source-code (Perl, C++, FORTRAN) is available via anonymous SVN at


LOOPP is a homology modeling server. It is based on a template detection algorithm learned by mathematical programming techniques that combines a large number of signals and significantly enhances typical detection capabilities (PSI-BLAST) by about 50 percent. It also uses a novel algorithm for alignment, and it finally builds atomically detailed models with Modeller (using the identified templates and our alignments of the target sequence into them). We use a combination of decision trees that constitute a “forest” to identify templates and assess the models. Each branch of the decision tree is a mathematical programming model and the confidence levels of the decision trees decrease as we move down the forest. The strength of the algorithm is in the very large training and test sets that we develop and use. The algorithm is fast and takes (at most) a few hours to build about 20 models per proteins.

LOOPP references:

Brinda Kizhakke Vallat, Jaroslaw Pillardy, Peter Majek, Jaroslaw Meller, Thomas Blom, BaoQiang Cao, and Ron Elber, "Building and assessing atomic models of proteins from structural templates: Learning and benchmarks", Proteins: Structure, Function, and Bioinformatics, 76:930-945 (2009).

Brinda Kizhakke Vallat, Jaroslaw Pillardy, and Ron Elber, "A template-finding algorithm and a comprehensive benchmark for homology modeling of proteins" , Proteins: Structure, Function, and Bioinformatics, 72:910-928 (2008).

Octavian Teodorescu, Tamara Galor, Jaroslaw Pillardy, and Ron Elber, "Enriching the sequence substitution matrix by structural information", Proteins: Structure, Function and Bioinformatics, 54:41-48(2004)

Jaroslaw Meller and Ron Elber, "Linear Optimization and a double Statistical Filter for protein threading protocols", Proteins, Structure, Function and Genetics, 45,241-261(2001)

Dror Tobi and Ron Elber, "Distance dependent, pair potential for protein folding: Results from linear optimization", Proteins, Structure Function and Genetics, 41, 40-16 (2000).

SABLE references:

R. Adamczak, A. Porollo and J. Meller,Combining Prediction of Secondary Structures and Solvent Accessibility in Proteins, Proteins: Structure, Function and Bioinformatics, 59(3): 467-75 (2005)

R. Adamczak, A. Porollo, J. Meller, Accurate Prediction of Solvent Accessibility Using Neural Networks Based Regression, Proteins: Structure, Function and Bioinformatics, 56(4):753-67 (2004)

A. Porollo, R. Adamczak, M. Wagner and J. Meller, Maximum Feasibility Approach for Consensus Classifiers: Applications to Protein Structure Prediction, CIRAS 2003 (conference proceedings).