The Proteomic Rationale and Approach

            Proteomics technology is like genomics in that it addresses large numbers of molecules simultaneously.  As stand alone technology, DNA and RNA based genomics often reveal little or no correlation to disease state, because there are so many regulatory steps between DNA and disease. These include, but are not limited to: imprinting, transcription, RNA splicing, RNA editing, translation, folding, post-translational modification, and heterochromatin, intracellular transport, secretion, functional interaction, and functional compensation.  Since proteomics is a protein based technology, it is several steps closer to the disease state.  Proteomics effectively detects post-translational modifications (see phosphoproteomics), leaving only intracellular transport, secretion, functional interaction, and functional compensation to be inferred.  If sub-cellular fractions are analyzed on separate gels intracellular localization can be determined.  If functional activity is controlled and separate gels run at different functional levels, then the functional state is known (see functional proteomics).  Also, kinetics information is available by running gels at different time points.  All of this data is easily merged, giving us many data points for each of our > 2000 protein spots.  These data points include: protein amount, post-translational state, and quantitative sub-cellular localization; These data points can be referenced to both their functional state and their kinetics information.  The data will become even more significant when combined with DNA and RNA based genomics, giving data points right though the entire biological system from DNA all the way to function.  All that is left is to infer how the function effects or causes the disease state.  Once this is done all of these data points can be analyzed for relation to the disease, to find the molecules that are contributing to the disease.

            Another advantage of protein based technology is that it is more easily translated into clinical use.  DNA and RNA based therapy and diagnosis are rare and mostly experimental, and probably will be for quite a while.  On the other hand, protein based therapy has been in use since before we knew what protein was. Some early examples are insulin extracted from animals and given as injections for diabetes, or antibodies to important proteins being used to diagnose disease.  Now recombinant proteins are in common use in hospitals everywhere.  Some proteomic technologies, like SELDI, are being used to find proteins that can be easily (relative to other approaches) translated to standard clinical lab tests.  They are also bringing closer the day when proteomics will be a speedy and powerful multifaceted clinical test.

          There are many different technical aspects to proteomics.  Preparative technologies range from 1D gels, 2D gels, Multidimensional chromatography, Mass Spectrometry, to Arrays, and beyond.  Each of these techniques has numerous different variations.  These techniques and their variations are applied singly, in tandem, and consecutively.  Identification and functional classification technologies include: Western Blotting, Specific Gel Stains, Mass Spectrometry, Database Mining, Computer Modeling/Simulation, and many others.  Of course there are many variations on each of these themes.Current technology can detect at most several thousand proteins and protein isoforms at once.   

Limitations of proteomics:  2D gels currently have the best available resolution in proteomics, about 2,500 spots per gel.  By alternating pI and % acrylamide it is possible to visualize up to 10,000 spots.  This is a giant leap forward from traditional SDS PAGE gels, but:  ~50,000 genes times an average of 6 splice isoforms per transcript, a variety of pre and post functional stages for most proteins, and a host of reversible and irreversible post translational modifications has lead to estimations of up to 40 million different proteins in humans, not counting the 100 million+ possible antibodies.  Almost all of these protiens will find their way into the blood at some time or in some disease state.  Additionally we'd like to know how much of each protein is present and where in or outside the cell it is.  Thus it is easy to see why current resolution is at least 7 orders of magnitude short of resolving all possible proteins.  However it is important to note that before the advent of proteomics only about 300 plasma proteins were known. Now through HUPO we have helped to cross-lab repeat an pilot phase experiment.  It confirmed more than triple that number raising the number to ~1000.  When this data is published clinical practitioners of every type will be looking through the list and finding new proteins that are clinically relevant to their specialty. Some of these will be developed into new tests and drugs to predict, diagnose and treat a wide variety of disease. In the meantime HUPO will be trying to go deeper into the proteome so that we can present 10,000, or 100,000 proteins within the next few years.

          There is no magic bullet in proteomics, only by using all of the currently available technologies can we get the most comprehensive view of a proteome.  This would be impossible in any one lab, although highly collaborative projects like those at HUPO, may eventually be able to leverage such resources.  Thus, each lab must consider what technologies are best suited to their own needs.  There are many factors in this decision, including: Resolution (number and kinds of proteins detectable), Sensitivity (quantitative range and accuracy), Reproducibility, and Throughput (relative rate of sample processing).  Perhaps most important are: Amount of money available, and access to experts that can make these technologies work.