Supplementary Material for

"Defining functional distance using manifold
embeddings of gene ontology annotations"

by G. Lerman and B. E. Shakhnovich

PNAS, June, 2007


Link to Paper

Link to Supplementary Text

Slides of Talk given at IMA & IPAM

Software and Supplementary Data (matlab): (all codes are documented)

The list of all 3306 protein domains can be found in here (webpage of B.E. Shakhnovich)

Processed Data File: In the attached file we indicate all GO annotations from April 2006 associated to each domain. We had only 3139 domains with meaningful annotations. For comparison we attach an older file (even though named new_data) with only 3032 domains (it applies older set of annotations from October 2005).

Creating the Unified Graph:
The attched file (create_graph_oct05.m) creates the unified graph and its appropriate weights and store relevant information in a format convenient for matlab use (all files are carefully documented). It calls the attached subprogram.
We have saved the resulted output of this process applied to our main file (individual_dom_matrices_process_component.out; attached above) in the attached file.

Computing the Distances:
Below are several codes for computing distances according to different kernels etc.
new_main_pseudo_lap.m uses the pseudoinverse of the graph laplacian and does not take an advantage of the sparse structure.
new_main_lle.m uses the LLE kernel and takes in part advantage of sparse computation.
new_main_diffusion_direct_powers.m uses Coifman and Lafon's kernels (using power of diffusion). The powers are applied here directly to the sparse matrices.
main_diffusion_sparse.m is similar, where instead of taking powers, the distances are computed via sparse eigendecomposition of the initial kernel (without power). It uses the following subprogrm.


The research described here has been supported by
NSF grant #0612608

Any opinions, findings and conclusions or recomendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Science Foundation (NSF).


Last Modified Thursday May 21, 2009
The views and opinions expressed in this page are strictly those of the page author. The contents of this page have not been reviewed or approved by the University of Minnesota.