Topic 1

2/24/2012 10:40:16 AM #

with a comprehensive experience in fundamentals and application for mass spectrometer,  I would like to express my concern that MS-based proteomics have been on a wrong track by spending too much in so-called shotgun-proteomics for protein identification since mid or late 1990's.   what shotgun-proteomics for protein identification contributed is that it does not solve any protein identification problems but keep creating avoidable problems to solve. As we saw in  about past 18 years,  someones keep getting funds for creating an avoidable problem, published the solution, reviewed by interest groups, to the problems they created, and then gets further funds.  that's a perfect way to get an endless funding at cost of tax payer money.

There is A shot-gun genomics, where  the method is chemically valid and solid for nt identification. But shotgun-proteomics is an invalid and groundless practice in  protein identification, this is because  Ion chemistry for MS/MS technique in peptide identification won't allow us to see all peptide bonds break at all let along the issues like co-eluting, dynamic range limitation.  MS vendors provides cutting-edge machines for the research, MS can tell m/z difference in sub-ppm level. But what did we got for a valid trace protein identification, please don't tell me abundant protein like abumin.

MS is a highly versatile and sensitivity detection tool, no doubt is it's a best tool for a protein or a peptide qunatitation, but it's not a tool for analyzing unknown mixture in highly denamic range. Spending money on protein purification and enrichment instead of shutgun approach.

We Americans all suffer from economy downturn, and Federal budget deficient. Please stop wasting any penny on shotgun protemics.
  

ming | Reply

3/15/2012 2:33:23 PM #

I think there is a need for wider use by the community of mass-tagging methods that permit quantification of changes in large sets of related proteins under different experimental conditions.  Sets of proteins could include phosphorylation of enzymes in regulatory pathways, surface recognition molecules associated with a particular stage of development, states of transcription factors that regulate defined gene sets, etc.  Changes in conditions could be different developmental stages, distinct hormonal signals, or time after a single signal.  These methods exist, but are still in limited use.  Training seminars or programs to widen technical knowledge about these methods would be useful.  Also, increased funding of MS instrumentation in shared research centers would be helpful.  These methods have the potential to surpass many standard quantification methods in both power and ease of use.

mbk | Reply

3/17/2012 8:02:55 AM #

So, speaking as a technology guy as well as a proteomics core director, proteomicist, informaticist and active grant reviewer and applicant, it seems that proteomics (read as the attitude of thought leaders and grant reviewers) has entered a period of relative experimental conservativism that works against order-of-magnitude improvements and emergence of disruptive technologies.  

We probably should look for guidance at the ongoing revolution in DNA sequencing and genomics, driven by the emergence of disruptive technologies and high quality informatics.  The first lesson is  that incremental improvements on current method are not going to get us there and obsession with obtaining high quality data "now" gets in the way of innovation.  NIH was not the hero here.  Don't forget that NHGRI, driven initially by Watson, opted to bet everything on gel-based technology in an insane race to get the first draft done, and chose to minimize investment in new technology development, even when many of us were proposing to develop next generation sequencing.  It may have seemed new, but the technology for Illumina sequencing is a combination of mid 80's to early 90's innovations.  The lack of risk-taking and focus on near-term deliverables likely held back progress in sequencing by five or ten years and squandered hundreds of millions of NIH dollars that went to enriching a single company.  The $1000 genome program came very late in the game, when proof-of-principle for sequencing by synthesis technologies was already in hand.  

So, I would argue that real money needs to be dedicated to high-risk (in the minds of the reviewers!) technology development.  For my part, my colleagues and I have been promising to take current mass specs and use computational tools to squeeze a couple orders of magnitude more protein IDs out of them, a couple of orders of magnitude faster than currently possible, yet no NIH investment has been offered.  Why?  Because it might not work?  It can't because we don't need these capabilities.  In turn, I am sure instrumentation-focused colleagues have proposed similarly disruptive approaches that would make current Orbitraps look like Model T's, but I bet they are similarly getting the cold shoulder.   Why not focus on impact rather than risk?   Is it really a problem that a new idea might not work so well that it obviates everything that has come before it?  

Importantly, as a model for innovation, at DARPA, program directors get fired if they fund things that work at too high a percentage.  How to do this at NIH?  Well, I would look at the history of dedicated NIH programs built around technology development.  NCI's IMAT program is as close to DARPA as NIH gets.  Yes, some stuff IMAT funds really isn't that creative and the really creative stuff doesn't always "work", but funding innovation is not easy.  The key thing is to have committed program staff and strong reviewers.  Simply gathering the typical collection of experts on something, particularly when the funding line is less than 20th percentile, is more likely to get you more of the same.

So, if money is going to be put into an RFA for "disruptive" proteomic technologies, the greatest challenge is at the steps of defining the FOA and then of performing review.  I would strongly suggest that an R21/R33 phased innovation award structure be used, with a panel review determining progress from R21 (two years, $300K) to R33 (three years, $1M), based on a progress report, with a goal of funding a quarter of the proposals at the R21 stage and then at the R33 transition.   Of course, irrespective of mechanism, there could be ten proposals submitted for every one that can be funded.  Picking the best with the greatest up-side will be a tough job.   If the wrong people are invited ad hoc for the panel, then there is no way to keep the funds from simply going into incremental, safe projects.   Building up a reliable reviewer base capable of looking at a wide range of technologies, whether in genomics, proteomics, imaging, or whatever, would be far better than bringing in a different group for each review.

Good luck!

Steve Kron at U Chicago | Reply