h a l f b a k e r y"It would work, if you can find alternatives to each of the steps involved in this process."
add, search, annotate, link, view, overview, recent, by name, random
news, help, about, links, report a problem
browse anonymously,
or get an account
and write.
register,
|
|
|
We've had the human genome sequence
for a long time, and most of the coding
regions (genes) have been identified.
This has accelerated research
enormously.
Previously, researchers interested in a
particular biochemical or developmental
pathway or disease might spend years
homing in
on the gene, which they would
then sequence. Today, better techniques
coupled with the complete genome
sequence often make it possible to
identify likely candidate genes in days or
weeks.
However, most of the proteins encoded
by
the genes are unknown and unstudied.
So, a research may find (by genetic
mapping, for instance) a chromosomal
region harbouring a likely "disease gene",
but the region will contain many dozens
of
genes, any one of which could be the
target. Or they may find a list of a
hundred genes which are over- or
under-
expressed in a given situation, any one
of
which may be the one they're looking for.
If we knew a bit more about each gene
(where it's expressed; what enzymatic
activity the protein has; which other
proteins it binds to; and so forth), it
would
be much, much easier to identify
candidate genes. For example, if you're
looking for the gene involved in a
metabolic defect, and you find a dozen
candidates of which two have enzymatic
roles likely to be involved in that
metabolic pathway, you immediately can
focus on those two genes instead of the
whole dozen.
There are "global" programs to catalogue
gene properties (eg, expression patterns;
effects of knocking out a homologous
gene in yeast, etc), but these are grand
surveys which give little insight into each
protein.
In contrast, a researcher working on a
particular topic will often spend years
studying a particular gene and the
protein
it encodes, and gain a very detailed
understanding of it.
I propose a proposal. There must be
about 27,462 new molecular biology PhD
students each year, worldwide. They all
need projects to work on. For one year,
and one year only, we assign each new
PhD
student a gene, which they will work on
for the duration of their project. If their
host lab is
interested in a particular gene or group
of
genes, then they get to work on one of
those, so everyone wins. If not, they can
pick one at random from the list of un-
analysed genes.
Over the course of three years, studying
any one protein is quite likely to turn up
something interesting (probably more
likely than most PhD projects). More
importantly, learning everything you can
about one gene and its protein (or
proteins
- since one gene may encode a number
of
variants) is an excellent way to develop
well-rounded skills, whatever the gene.
In three years' time, then, we'll have
27,462 PhD theses, each with an in-
depth
analysis of one gene (and its
ramifications). These theses would be
made available on-line (as I believe all
theses should, though few are), linked
directly to the many databases
cataloguing
the human genome. In one swell foop,
we'd accelerate research by a significant
step.
Folding@home
http://folding.stanford.edu/
Uses home computers and PS3s to run detailed analysis of protein folding. [mecotterill, Jul 14 2008]
Please log in.
If you're not logged in,
you can see what this page
looks like, but you will
not be able to add anything.
Annotation:
|
| |
Are there many stand-alone genes? I'd have guessed they work, at least in part, with other genes to produce a given outcome. Fault tolerance, if you will. |
|
| |
I get my jeans out of the Levis's
catalogue... how annoying is that for an
annotation? |
|
| |
I believe PhDs write 'dissertations'. Anyways, just to clarify, this idea is simply to adopt a gene centric year of research projects for biochemists in order to sustain a library of information relating to each gene? And who is going to pay for it? Who is interested in the gene that encodes for the protein that regulates the gene that encodes for the protein that regulates the gene... I hope you get the point. Nobody is interested in that little gene, however noble it's cause, and nobody is going to pay for the research. |
|
| |
// nobody is going to pay for the research. // |
|
| |
Wrong. Tell "The Government" that it is "a potential weapon". Then everyone pays for it .... |
|
| |
//I propose a proposal//
hahah
//There must be about 27,462 new molecular biology PhD students each year//
haha
I like you.
+ for good phrasing |
|
| |
//Are there many stand-alone genes?//
No, there are none. However, many
researchers focuss on a particular
protein, including its interactions with
other proteins. Such studies inevitably
overlap, which is all to the good. |
|
| |
//I believe PhDs write 'dissertations'.//
Here, they're usually called theses. |
|
| |
//just to clarify, this idea is simply to
adopt a gene centric year of research
projects...// Yes, except it's three years
(all PhDs starting in a given year; a PhD
normally takes three years in the UK, or
longer in the US where it includes some
preliminary work). |
|
| |
//who is going to pay for it? Who is
interested...Nobody is interested in that
little gene, however noble it's cause,
and nobody is going to pay for the
research.// That would have been the
attitude 20 years ago, but not now.
First off, almost all genes are
interesting if you track them down (this
one is a dull copper transporter; but it
turns out that a mutant form of it
transports chromium and kills people
by chromium accumulation in cells; this
one is just a foetal globin gene, but it
turns out that it's also expressed in
adult hypothalamus and helps to
maintain yada yada); there are very few
dull genes, partly because of the
synergies that phoenix mentioned. |
|
| |
Second, we already pour huge amounts
of money into researching specific
genes (or their products), in the course
of following a particular biological
story. It's very much like the situation
in genomics pre 1990's - every
researcher went and found and
sequenced the gene they were
interested in. In the end, it was orders
of magnitude more efficient to just
sequence the whole damn genome and
lay it out on a plate, even though
everyone thought most of it would be
dull (and much of it is). Likewise, there
are
proteomics projects already underway
and heavily funded to categorize certain
aspects of every protein in an
industrialized manner. However, this
does not give you the depth of
knowledge and background that you
will get from having one person
focussing on one protein for three
years. |
|
| |
Don't think that this would be drudgery.
Almost any research topic is fun and, in
the long run, as good a training as any
other. You'd have a generation of well-
trained PhDs with diverse skills and,
into the bargain, detailed personal
information on every protein. |
|
| |
// detailed personal information on every protein. // |
|
| |
"The protein police are watching YOU ....." |
|
| |
But that's how it is. A real
understanding of a disease often comes
from knowing the habits and
behaviours of the relevant proteins -
the key players - inside out. If you're
looking for a key oncogene, and you
have a candidate region containing two
dozen genes, it takes an understanding
of those proteins to be able to say "well,
that one's only a cytoskeletal
component; but we know that, oddly, it
binds cyclic AMP which is one of the
factors that can accelerate growth of
these cells in culture..." |
|
| |
I accept the goodness of the idea, even though it still strikes me as grossly redundant. For starters, can the whole process be further automated (like the genome)? Let's train a robot to track each and every protein with florescent tags and log the data, and then just let a handful of students hack away at this insurmountable task for a few years. Also, you keep going on about how well trained everyone is gonna be, but once it's done, it's done. They'll be masters at a game nobody plays anymore. |
|
| |
//Let's train a robot to track each and
every protein with florescent tags and
log the data, and then just let a handful
of students hack away at this
insurmountable task for a few years.//
This is precisely what's done at present
(in flavour if not in detail).
It's too shallow - the robotic part (and
the mass array screens, and the mass
everything else) is
actually trivial; the latter part is
intensely human and requires huge
collective effort. What you're saying is
equivalent to saying
"let's build a robot to collect forensic
evidence and then we just need a few
detectives to analyse it." |
|
| |
// They'll be masters at a game nobody
plays anymore.// No, they won't.
There isn't a "game" of analysing
protein function - not in a deep sense.
Depending on the protein, it will
require a huge diversity of skills. For
instance, you clone the gene (widely
useful skill) and express the protein
(itself a black art and widely useful),
then maybe use it in a "pull-down"
reaction to identify what it might bind
to (which is itself a whole host of other
sub-skills, all of general utility), then
try to identify those proteins by peptide
fingerprint mass-spec (again, not made
redundant by the goals of this project),
then maybe you find that it binds
proteins X and Y, but that X also binds
Z, so maybe your protein shuts down
the activation of Z which is involved
in....etc etc. You will need the entire
spectrum of molecular biology tools to
really understand a single protein in
detail; almost all of those skills have
wider applications. Plus, we're talking
about documenting the normal proteins
in human cells - that leaves a few other
species. |
|
| |
Yes. This needs to be done, immediately. I still think funding is your enemy, though. You need a government 'human proteome project' to collect the bills and dole them out to the schools. In fact, you might want to consider this as a new title. Don't worry about rights, I've already lost them all... |
|
| |
Oh wait, your title is funny. Nevermind ;) |
|
| |
I think a government program (and
maybe a 'governments' program - let's
not forget the USA) would work. But
there is an alternative. Just let the
government subsidise 25% of the costs
of any PhD which is part of this project. |
|
| |
Most research groups would leap at the
chance to receive some additional
support, and in most cases could easily
dovetail a protein-targeted project into
their overall research strategy. It's
largely a matter of co-ordination and a
common goal. |
|
| |
Another reason for doing it in one big
cohort is that, if it drags on, you'll start
to get "formula theses" in a standard
format, as each student follows the
path of his predecessor; we want to
keep the novelty of approach. |
|
| |
Good question. You need to get above
a threshold of competence, inasmuch
as you don't want to add disinformation
to the databases. Beyond that, almost
any human analysis - even if it's no
more than a good trawl of the existing
literature and some clever
bioinformatics - can help. |
|
| |
But, really, I think you want a spectrum
of research to give a comprehensive
biography of the protein. I suspect a
PhD is about the right length for this,
though a Masters might cover it. |
|
| |
Incidentally, I'm voting [-] on this
because I can see some drawbacks and I
don't think it merits unanimous
support. |
|
| |
April 1 (Bloomberg) -- Harvard College rejected 93 percent of its applicants for next fall's class, the highest rate in its 372-year history. |
|
| |
Harvard admitted 1,948 of 27,462 applicants, the Cambridge, Massachusetts, school said in a statement today. That's a 5.3 percent drop from a year ago, when the college, the undergraduate arm of Harvard University, admitted 2,058 from a pool of 22,955. |
|
| |
Presumably, the ones who didn't make it go to work in a jean factory or something. |
|
| |
If you Pile them High and Deep enough, they can do anything. The scientific method: have your grad student do it. |
|
| |
I think the genes shouldn't just be assigned at random, though, since there are more prized targets than others. |
|
| |
Perhaps you should do something like a sports league draft, letting the top grad students get the first pick on projects they would like to work on. |
|
| |
Perhaps some big-name institution could fund this, for maybe 100 students or so, working on 100 Open-reading frames of unknown function. If students agree to the stipulations of the draft, and they are accepted based on academic merit, then they get the moneys. |
|
| |
So, not a new Katie Melua song title? |
|
| |
I'm having trouble trying to be positive about the idea [MB], but I'm having even more trouble trying to articulate my criticisms without answering them myself. I still remain unconvinced though - this could be a two pipe problem. |
|
| |
In the meantime, where you going to get 27,462 half-decent PhD students from? |
|
| |
Here's why I ask - there are 126 universities in the UK (just choosing UK because I know the number - we can scale this up for an international reckoning). I don't know how many of them have a half-decent biology department (that has genetic research in its books). Let's be generous and say 75. of that 75, 5 might have big departments and be renowned for genetics research. Could they have 10 PhD students?? I don't know... Anyway, the rest could have 5 PhDs in gene-related research. That's 400 PhDs in the UK per year... |
|
| |
Does that sound plausible? |
|
| |
Plus there is that fact that some (A lot?!) of PhD students' progress in the first year can be unpredictable. |
|
| |
{Just noticed - I take it that molecular biology is a superset of genetics, so my "figures" are subject to debate.} |
|
| |
"Oh my God! - this gene causes people to post complex, humourous ideas lacking rigour and practicality on internet fora!" |
|
| |
//I can haz jene reseach!// marked for
tagline. |
|
| |
//where you going to get 27,462 half-
decent PhD students from?// Yes, I did
sort of pluck that idea out of thin air.
My lab recruits maybe 50 a year, so I
was guessing 1-2000 across the UK in
decent departments, then multiply up
by the US, Japan, China (who are pretty
good), France, Germany, Malaysia....
But it's a soft number. |
|
| |
//Get the post-docs to do this// The
problem is that most post-docs have
formed their ideas about what they want
to research on, and are unlikely to be
swayed. Students are more flexible, and
would do a good job whilst acquiring a
good skills set. |
|
| |
Shovelling fries in McBurgers is the usual career path ..... |
|
| |
//to the point where many will not
complete a PhD.// I don't know what the
completion rate is elsewhere, but in the UK
I'd guess it's
well over 90% (I'm speaking only for
molecular biology, and only for the labs I
know). |
|
| |
I think overall in the UK it's about 50% - or
it was when I did my PhD. And my
grumble about non-completing PhD
students: I get really annoyed with people
who tell me "Oh yes, I nearly got a PhD - I
had about 6 months of writing-up left and
then I went and got a job". They have
completely missed the point: The last six
months of writing-up is the hard bit. |
|
| |
Plus, if you drop out with "only the writing
up" left to do, it's a pretty good indicator
that you didn't have anything worth
writing up. |
|
| |
I'm surprised it's as high as 50%, though
presumably that varies across discisplines? |
|
| |
// that varies across discisplines//
No, English doctorates have a much lower drop-out rate. |
|
| |
Yes, fair points. Not sure that the getting
scooped issue is a particular concern - the
same's true whatever field you pick. |
|
| |