Structure Prediction Protein Design Biclustering + cMonkey Network Inference Publications          Links + Collaborations People                   Opportunities             Software + Code Teaching            


WCG Status Update

WCG Post

why HPF2 is important: disease relevance
[Nov 10, 2006 6:44:40 PM]

warning the following has only been edited by sleep deprived people ... my wife and I just had a kid ... so that's areason for my lapse in the forum ... but not an excuse ...

Please take a look at our collaberators websites, and see for yourself what they are doing, for those of you who don't want to read all this below, tha basic idea is that the higher res structure predictions will be part of how we provide bioinformatics support (along with other bonneau-lab activities) to these labs.

Disease relevant areas of focus for HPF and HPF2: Three Specific Collaborations Centered on Annotating Disease Related Proteins enabled by HPF2.
Here we detail three of our main efforts to get the information resulting from the grid into the hands of groups working on disease. In each case the our lab has begun to put together tools and databases to suit the specific needs of these three research efforts. In general HPF and HPF2 are providing vital bioinformatic support, helping these groups to understand the structure and function of proteins central to their research efforts. We encourage you to explore the websites of each of these three groups to learn for yourself more about their efforts.
These collaborations will inform our development of tools for integration of organism-specific information with our structure-prediction-derived information. We describe the efforts ongoing in these three labs and then go on, in each case, to describe how this project will interface with these efforts in terms of anticipated development of tools and prioritization of aims and development required by each specific effort. This work is being carried out in collaboration with Lars Malmstrom, in the Goodlett lab (Medicinal Chemistry Dept., University of Washington). In each case we describe a collaboration that explains why we have asked you to fold these proteins. Each collaboration corresponds to a set of proteins in the HPF2 queue.

Protein Set #1: Malaria proteins.
Collaborator: Patrick Duffy, Paul Shannon, Seattle Biomedical Research Institute, Seattle,
The Duffy lab aims to create a pregnancy malaria vaccine. In 2003, Dr. Duffy and a consortium of laboratories launched the Pregnancy Malaria Initiative to identify the necessary antigens for a malaria vaccine to protect women during pregnancy. They are using bioinformatics, microarray and proteomics tools to characterize the distinct features of these parasites and evaluate parasite surface proteins that may be developed as pregnancy malaria vaccines. The proteins identified by the consortium are now being assessed by the Human Proteome Folding project (HPF2), in order to understand their function and structure so that vaccine designs can be improved. Using the paradigm established in their studies of pregnancy malaria, the Duffy lab has also launched a program to develop vaccines against severe childhood malaria. With support from the Grand Challenges in Global Health (GCGH) Program, an international consortium led by the Duffy lab is now trying studying the immune responses that protect African children from severe malaria. African children may only suffer one or two episodes of severe malaria before developing resistance, and earlier studies showed that antibody purified from the serum of immune adult Africans could cure young children with malaria. The consortium is thus identifying parasite proteins that may be targeted by protective antibody, as a key step in developing vaccines for children. As part of the Human Proteome Folding project we are working with the Duffy Lab to dramatically improve their ability to annotate many of proteins they have recently found to be important to Plasmodium and its specific interactions with its host.

Protein Set #2: human cancer biomarkers of unknown structure and function.
Collaborator: Leroy Hood, Nathan Price, Institute for Systems Biology, Seattle,
Multiple groups involved at the Institute for Systems Biology are currently involved in a coordinated effort to characterize biomarkers that can be used for early diagnosis and sub-classification of Human cancers. In particular specific efforts are underway in the Laboratory of Leroy Hood to find prostrate, bladder and ovarian cancer biomarkers. This effort coordinates proteomic, microarray, pathology, and bioinformatics efforts into a overall effort to determine reliable and readily assayable predictors that can be used as markers for diagnosis and selection between alternate therapeutic/intervention regimes. The Bonneau and the worldcommunity grid lab are involved in the functional annotation of putative proteins and proteins of unknown function found in these studies. To date several hundred putative biomarkers of unknown function have been prioritized, and are being processed along with the other sets of proteins described above, on the world community grid. Along with Nathan Price, at the ISB in Seattle, the Bonneau lab has been applying structure based annotaiton to elucidate the structure/function of putative biomarkers discovered using these genome-wide screens.

Protein Set #3: Gram-negative pathogens.
Collaborator: David Goodlett; University of Washington, Seattle:
: Dave Goodlett's laboratory has used Francisella tularensis subspecies novicida (strain U112), a murine pathogen, as a model to study virulence in Francisella tularensis, a human pathogen. Both organisms are extremely virulent to their respective hosts causing high morbidity/mortality if left untreated by antibiotics. Like familiar Gram-negative bacteria, such as Yersinia pestis, F. tularensis is a class A agent. Our (Bonneau lab) primary roll, so far, has been to carry out genome annotation for the most difficult proteins in these gram-negative pathogens using our structure-inclusive pipeline.
The Goodlett lab was able to verify that many proteins in these genomes with no homology to other genomes are actively expressed at different conditions, increasing our interest in applying the described methods to these genomes. U112 strain genome consists of a single circular chromosome of 1,910,031 bp (32.48% GC) compared to previously sequenced F. tularensis SchuS4 strain with 1,892,819 bp; ~ 350 Kb of genome is unique to U112. Among other findings U112 has a single copy of the pathogenicity island sequences versus two in SchuS4 and significantly more complete metabolic pathways. Database search of tandem mass spectra detected 65.4% of predicted genes as expressed proteins with false discovery rate of < 0.01% using a mock database. In combination with genome annotation ~ 80% of genes are predicted to be expressed. Of predicted genes ~ 30% had no homology to genes encoding proteins of known function preventing corroboration of their authenticity. However, observation of gene products validated authenticity of > 50% of these hypothetical genes representing 23.2% of all expressed genes; no pseudogenes observed. Finally, we are using Rosetta de novo, fold recognition and homology-modeling to predict structure and infer function for many genes of unknown function.

more in a few weeks,

monkey monkey monkey monkey monkey monkey monkey monkey monkey monkey
[Edit 1 times, last edit by rbonneau at Nov 10, 2006 6:46:09 PM]