Pages:
1
2 |
ampakine
Harmless
Posts: 25
Registered: 29-4-2011
Member Is Offline
Mood: No Mood
|
|
Programmer/chemist
I'm studying chemistry in college but I seem to have more of a passion for programming and computer related things so I'm gaining just as much
programming and computer skills as I am chemistry skills these days. I want to get into a chemistry/pharmacology related field but I'm wondering if my
programming skills will come in useful. I'm getting lots of ideas for ways for developing chemistry software and technology and I don't just mean
software for making models such as ChemDraw, I mean developing hardware and software for automating chemistry processes. One example off the top of my
head: I could develop a machine that will automatically determine the solubility of a compound in a wide range of different compounds at different
temperatures. If I'm not mistaken this kinda thing is usually done manually by a chemist. With the right software a machine could do it far more
effectively than a human.
Any other programmers/technophiliacs here?
|
|
phlogiston
International Hazard
Posts: 1379
Registered: 26-4-2008
Location: Neon Thorium Erbium Lanthanum Neodymium Sulphur
Member Is Offline
Mood: pyrophoric
|
|
In my daily work (as a biochemist) I get lots of ideas for automation too. However, in my opinion, the critical factor is: how big is the market for
your machine. Either the machine needs to cost a fortune for it to pay your bills, or you have to sell a lot of them. AFAIK, there are not that many
occasions when you need to establish the solubility accurately.
Some of the tasks I think are worth automating (ie. I would buy a 10,000+ Euro machine to automate the job):
1. Calibrating pipettes
2. Western blotting
3. transfering samples from vials to 96 well plates
-----
"If a rocket goes up, who cares where it comes down, that's not my concern said Wernher von Braun" - Tom Lehrer
|
|
bahamuth
Hazard to Others
Posts: 384
Registered: 3-11-2009
Location: Norway
Member Is Offline
Mood: Under stimulated
|
|
If you like programming and like biology you should aim for bioinformatics.
Pay is good as I understood it, e.g. on projects one might get between $90-300K a year in Norway. You would earn more if you work freelance.
Developing hardware is expensive (my first biotech patent filed two weeks ago), plus there is the thing about patenting your invention, and reading
prior art and such as not to be sued.
But yeah, you should get into it.
PS. One tip, next generation sequencing is the big thing and will continue to be so for atleast a decade IMO.
Any sufficiently advanced technology is indistinguishable from magic.
|
|
phlogiston
International Hazard
Posts: 1379
Registered: 26-4-2008
Location: Neon Thorium Erbium Lanthanum Neodymium Sulphur
Member Is Offline
Mood: pyrophoric
|
|
That kind of money....probably not as a postdoc, right? Industry?
-----
"If a rocket goes up, who cares where it comes down, that's not my concern said Wernher von Braun" - Tom Lehrer
|
|
bahamuth
Hazard to Others
Posts: 384
Registered: 3-11-2009
Location: Norway
Member Is Offline
Mood: Under stimulated
|
|
Well, a postdoc position in a government place e.g. like the university I work wouldn't get you more than lowest pay they are allowed to give you
(those cheap bastards..) which would be less than $90K. We lack a fulltime bioinformatician in our group and so, some use their presious off work time
to learn the skill. The reason we lack one is that nobody would take a doctorate and be paid that low, and the reason the administratives gave for the
low pay is that none could earn more than their schooling as not to "upset" the balance even if their skill was unique. So yes, one would have to go
into the private sector to get to the real money..
Any sufficiently advanced technology is indistinguishable from magic.
|
|
Mildronate
Hazard to Others
Posts: 428
Registered: 12-9-2009
Member Is Offline
Mood: Ruido sintetico
|
|
Quote: Originally posted by phlogiston | In my daily work (as a biochemist) I get lots of ideas for automation too. However, in my opinion, the critical factor is: how big is the market for
your machine. Either the machine needs to cost a fortune for it to pay your bills, or you have to sell a lot of them. AFAIK, there are not that many
occasions when you need to establish the solubility accurately.
Some of the tasks I think are worth automating (ie. I would buy a 10,000+ Euro machine to automate the job):
1. Calibrating pipettes
2. Western blotting
3. transfering samples from vials to 96 well plates |
To my mind humands need to forget about volume measuring at all!
|
|
resveratrol
Hazard to Self
Posts: 65
Registered: 6-11-2010
Member Is Offline
Mood: No Mood
|
|
Ampakine,
If you attend a university, you should consider taking a graduate course in Computational Chemistry (i don't think they teach it undergrad). This
will give you a feel as to what to expect in the emerging computational/programming aspects of chemistry. It's funny you mention developing
"machines" or instruments that determine the solubility of compounds in a wide variety of solvents, as this is getting more into the engineering
aspect. There are actually clients available that can compute this information for you. You just have to learn how to use them (and often times
you're using several to dozens of different programs in conjunction to the client you use to build your molecules). I found the coding to be very
frustrating when I was taking this class, because of the amount of errors you generate in trying to pump out calculations (which can take days to run
even in a university high-performance computing center). But it's also very fascinating to see what's possible with this software, and what people
have done with it. And I see you mentioned pharmacology - this is actually very relevent to modern-drug development/pharmacology as well.
|
|
zoombafu
Hazard to Others
Posts: 255
Registered: 21-11-2011
Location: U.S.
Member Is Offline
Mood: sciencey
|
|
If you want to explore developing hardware, i highly reccomend buying an arduino kit Arduino Website. Ive got one and I have set it up with multiple sensors so I can measure temp, pH, and I can even measure O2 percentage in the
air, etc.
|
|
monglober
Harmless
Posts: 2
Registered: 15-11-2012
Member Is Offline
Mood: No Mood
|
|
The answer:
automation is already a huge part of big-pharma.
I don't just mean autosamplers/injectors or the robotic SP synthesis of DNA/RNA or peptides....
Combinatorial chemistry.... its huge...... its broken down into software and hardware.
the software is bioinformatics + computationa/theoretical chemistry/biochemistry + computer science (+pure math)
typically a big push is to find ways to simulate ab-initio/semi-empirical or semi-classical (for MD) models using whatever algorithm/technique can
achieve it....
the goal is to minimize CPU time without unaccetable error.
Hardware is genuinely a fast, automated, testing procedure..... thousands of chemicals can be synthesized using microfluidics (nanomolar quantities)
and robotics and automatically shot into HPLC/MS or even NMR in a fully automated process.
typically this does not mean multi-step synthesis of an entire drug molecule.... it means testing 1 step of a proposed synthesis...
when it is used to test the actual drug molecule itself, they will often synthesize thousands of different chemicals with slightly different
structures (eg: 2 carbon vs 3 carbon chain.... 5 carbon vs 6 carbon ring... etc), then pass those potential compounds through to a genuine
biological/biochemical target (such as a lab-on-chip system) which will give a measurable response depending on the binding efficiency.
the reality is that the field is full of proprietary technology and trade secrets..... some of it is COTS (commercial off the shelf) but most is
not.
|
|
Chemosynthetic
Harmless
Posts: 19
Registered: 12-11-2013
Member Is Offline
Mood: No Mood
|
|
Not to resurrect an old thread, but it is near the top, so I will respond. I do some computational aspects within drug discovery (including QSAR, data
mining and drug docking, primarily), and know a couple others in pharmacology who do much heavier programming than I. While my experience in purely
chemical modeling is shallow, I have found much more interaction between in silico and wet pharmacology since transitioning from just chemical
synthesis. Neuroscience modeling and some proteomic aspects of pharmacology seem to largely rely on custom single purpose programs despite several
great open source suites (ex. NEURON, ADT). I have also found scripts useful in genome annotating, though you can use many publicly accessible
resources for that type of work as well, like BLAST, CLUSTAL, etc.
|
|
packetforger
Harmless
Posts: 48
Registered: 21-2-2014
Member Is Offline
Mood: Condensing
|
|
zoombafu: Have you any resources on pH probes and interfacing software for the Arduino? Is there sensors one can use for density of a solution?
I recently was thinking to fit similar sensors to one of my homebrewing containers, see if I could graph pH changes, temp changes in solution over
time (along with, if possible, density changes and the likes) but did not know where to start, having never done any work with the Arduino itself
before (my programming background being networked applications and kernel modules for Linux). Any advice would be greatly appreciated!
As for OP, the computational chemistry/bioinformatics side of things is supposed to be where a lot of the "serious money" is these days.
|
|
packetforger
Harmless
Posts: 48
Registered: 21-2-2014
Member Is Offline
Mood: Condensing
|
|
zoombafu: Have you any resources on pH probes and interfacing software for the Arduino? Is there sensors one can use for density of a solution?
I recently was thinking to fit similar sensors to one of my homebrewing containers, see if I could graph pH changes, temp changes in solution over
time (along with, if possible, density changes and the likes) but did not know where to start, having never done any work with the Arduino itself
before (my programming background being networked applications and kernel modules for Linux). Any advice would be greatly appreciated!
As for OP, the computational chemistry/bioinformatics side of things is supposed to be where a lot of the "serious money" is these days.
|
|
flashmanc
Unregistered
Posts: N/A
Registered: N/A
Member Is Offline
|
|
Nice one
Perhaps I'm wrong, but in my opinion programming and chemistry have nothing to do with each other. If you are interested in both sciences, perhaps
that means that you have a drive passion for learning difficult things.
|
|
The Volatile Chemist
International Hazard
Posts: 1981
Registered: 22-3-2014
Location: 'Stil' in the lab...
Member Is Offline
Mood: Copious
|
|
How about in-organic chemistry? Are there any suggestions as to what kinds of programs would be useful for even amateur chemists?
|
|
CrimpJiggler
Hazard to Self
Posts: 75
Registered: 1-9-2011
Member Is Offline
Mood: No Mood
|
|
I do some programming myself too, I see massive potential in the area of computational chemistry, especially since its only gonna become more and more
commonly used as computers become more powerful. As someone above mentioned, its mighty useful for QSARs, most chemists aren't programmers so they
don't know how to write a script that will automatically generate a thousand Gaussian input files containing slightly modified z-matrices and
compiling it all into a batch file that Gaussian will run automatically. Then on top of that, you can do data mining on it, and find correlations that
would be difficult for a human to spot.
The Volatile Chemist: Yeah, I bet there are plenty of applications. You can predict reaction pathways with Gaussian, so that'd be one use for an
amateur chemist. I haven't got into computational chem much yet, I don't know how to predict things like solubilities, not sure if the technology is
even there yet but that would be very useful for amateur chemists (any chemists for that matter). Someone above mentioned an arduino kit, I clicked
the link and that looks awesome. I don't know anything about electronics, I wanna get into it because I would be able to write programs to control the
machines. You could build something like an automatic columning machine, once it detects that a product is coming out of the column, it waits for it
to elute fully, then switches container. I'm getting a bit carried away here though, I have no idea how I could do that.
Quote: Originally posted by flashmanc | Perhaps I'm wrong, but in my opinion programming and chemistry have nothing to do with each other. If you are interested in both sciences, perhaps
that means that you have a drive passion for learning difficult things. |
If you take up programming, you'll see it differently. The physical world is kind of like a program, everything has an algorithm to it. Thats what
QSARs is all about, finding those algorithms/correlations. Chemistry and programming are two separate fields for now, but in the future they'll be a
lot more interlinked. Computational chemistry will make a lot of practical experiments obsolete, which is a good thing because it'll mean less wasting
of resources, and in medicinal chemistry, it'll mean less need for animal testing. The main issue now is that it takes such a long time to run
calculations on big molecules, but that will change. Look at how far computing technology has advanced in the past 10 years.
[Edited on 28-4-2014 by CrimpJiggler]
|
|
woelen
Super Administrator
Posts: 8020
Registered: 20-8-2005
Location: Netherlands
Member Is Offline
Mood: interested
|
|
Quote: Originally posted by flashmanc | Perhaps I'm wrong, but in my opinion programming and chemistry have nothing to do with each other. If you are interested in both sciences, perhaps
that means that you have a drive passion for learning difficult things. |
This simply is not true. Programming (and more general: data aquisition and control) becomes more and more important in chemistry. I hardly can
imagine a single chemical plant which is not leaning heavily on ICT, also for its core processes. Control and monitoring software makes plants more
reliable, makes it easier to precisely control amounts of reagents to be mixed and allows an operator to peek into the process at great depth.
And what to think of all modern measurement and analysis equipment, used by chemists. These things really are full of software, for converting raw
measurement data to intelligible data, presenting results on screens, controlling the measurement itself.
Software also becomes more and more important in research, as mentioned by the previous poster. Analysis of large molecules can be done by means of
advanced molecular models which are simulated by software. This requires tremendous amounts of processing power, but computers still are becoming
faster rapidly and at the moment there also is a development towards more and more parallellism (multiple cores) and at the software level the use of
multiple cores is exploited more and more.
|
|
quantumchromodynamics
Hazard to Self
Posts: 67
Registered: 25-9-2013
Location: with much determination, nowhere in particluar
Member Is Offline
Mood: tired but still trying
|
|
programming and chemistry
I lost my baby teeth coding MAC drivers for industrial imaging scanners. I have written knowledge management systems, computer vision systems, built
military weapons controllers, and everything else computer related under the sun, for over 30 years. IMHO programming is vastly interesting and very
much related to the scientific method. Your thoughts must be absolutely precise to program well. Programming feels mathematical in the states of
concentration it requires. Programming machine to machine, the internet of things, the recent influx of interesting tiny computers, cell phones,
raspberry pi's, aurdrino's, etc, is changing the world. I guarantee a foundation in software development will make you more effective at any technical
or scientific career that you choose. I guarantee that writing quality code makes you a better scientist, and therefore, a better chemist.
|
|
Dagny
Harmless
Posts: 2
Registered: 15-5-2014
Location: Colorado, USA
Member Is Offline
Mood: No Mood
|
|
Chemistry and computing certainly have strong and interesting interconnections (my entire job/career is based on this) Quickly surveying this connection-- a vast array of computational methods in
chemistry and biochemistry have been developed for creating 'realistic' models of molecules, to be able to predict reactivity or properties.
Experimental chemistry results provide feedback to develop these models, and in turn, once the models are refined sufficiently, they're able to make
high-accuracy predictions of new molecules or reactions (as long as the inherent chemistry of the 'new' system is still appropriate for the set of
approximations being used in the model). Also, when interesting phenomena in chemistry or biochemisty are observed, it is often hard to unravel a
molecular-level explanation in the lab, but through these precisely-controlled computational models, we can get a picture of what may be happening and
improve our knowledge of the basic sciences.
Computational chemistry ends up being a very powerful tool for its predictive or explanatory power, and also can save a lot of money in academic or
industrial research settings, by just burning some CPU hours, to reduce the number of molecules or reactions an experimentalist may need or want to
test in the lab (reducing both raw material costs, and of course human *time*)
Opportunities for careers in this area are intriguingly diverse, from predicting new catalysts, new material properties, understanding basic organic
chemistry reactivity, improving photovoltaics, understanding enzyme mechanisms, ...( I'll refrain from continuing an overzealous list-- it's truly an
endless list of fascinating applications, to fill far too many lifetimes )
|
|
The Volatile Chemist
International Hazard
Posts: 1981
Registered: 22-3-2014
Location: 'Stil' in the lab...
Member Is Offline
Mood: Copious
|
|
Quote: Originally posted by Dagny | Chemistry and computing certainly have strong and interesting interconnections (my entire job/career is based on this) Quickly surveying this connection-- a vast array of computational methods in
chemistry and biochemistry have been developed for creating 'realistic' models of molecules, to be able to predict reactivity or properties.
Experimental chemistry results provide feedback to develop these models, and in turn, once the models are refined sufficiently, they're able to make
high-accuracy predictions of new molecules or reactions (as long as the inherent chemistry of the 'new' system is still appropriate for the set of
approximations being used in the model). Also, when interesting phenomena in chemistry or biochemisty are observed, it is often hard to unravel a
molecular-level explanation in the lab, but through these precisely-controlled computational models, we can get a picture of what may be happening and
improve our knowledge of the basic sciences.
Computational chemistry ends up being a very powerful tool for its predictive or explanatory power, and also can save a lot of money in academic or
industrial research settings, by just burning some CPU hours, to reduce the number of molecules or reactions an experimentalist may need or want to
test in the lab (reducing both raw material costs, and of course human *time*)
Opportunities for careers in this area are intriguingly diverse, from predicting new catalysts, new material properties, understanding basic organic
chemistry reactivity, improving photovoltaics, understanding enzyme mechanisms, ...( I'll refrain from continuing an overzealous list-- it's truly an
endless list of fascinating applications, to fill far too many lifetimes )
|
The Main problem for me is that I'm a programmer (Per-say, I haven't taken real classes, just read), and an uneducated amateur Chemist. I don't know
how to data mine chemistry subjects, I know nothing of molecular theories, etc. I need someone to tell me what can be done and how to do it, or at
least give me a book on it I'd love to write programs for chemistry, but where
can I find books on predicting catalyst properties in chemicals? I don't care how complex the subjects are, but I am yet to find an explanation.
|
|
rskennymore
Harmless
Posts: 12
Registered: 10-3-2014
Member Is Offline
Mood: Eutectic
|
|
Quote: Originally posted by The Volatile Chemist |
The Main problem for me is that I'm a programmer (Per-say, I haven't taken real classes, just read), and an uneducated amateur Chemist. I don't know
how to data mine chemistry subjects, I know nothing of molecular theories, etc. I need someone to tell me what can be done and how to do it, or at
least give me a book on it I'd love to write programs for chemistry, but where
can I find books on predicting catalyst properties in chemicals? I don't care how complex the subjects are, but I am yet to find an explanation.
|
I'm more or less in the same boat here. I'm a software engineer by trade, I've been developing java for the better part of 5 years now. I've been
interested in chemistry for most of my life but have never taken the plunge into trying to understand it at a fundamental level until recently. I've
sorta just started writing a molecular geometry engine and am fleshing it out as I learn more about how atoms interact. Mostly reading wiki articles
and the vast sea of pages you can find from google searching. I think trying to understand how to predict catalyst properties without understanding
atomic models would be putting the carriage before the horse, that might just be me though.
|
|
smaerd
International Hazard
Posts: 1262
Registered: 23-1-2010
Member Is Offline
Mood: hmm...
|
|
Quantum chemistry calculations are no joke. This is where not only must one be a good programmer writing in a low-level environment for speed, one
must be a good mathematician, and physicist/physical chemist. The things people do to turn abhorrent calculations into bit shifts and things are
absolutely mind-bending(to me but I'm a laymen). Even the 'fast inverse' square root that was used in the quake 3 engine took my breathe away the
first time I saw it hahaha(http://betterexplained.com/articles/understanding-quakes-fas...). I have written a few applications with my background in 3-D programming related
to the quantum nature of single atomic orbitals. It wasn't difficult but it did take time to get everything right(3-4 days), and on that level well,
it wasn't useful more demonstrative.
There are tons of resources for this kind of information.
Start with things like - http://vergil.chemistry.gatech.edu/courses/chem6485/pdf/Elec...
Move to things about numerical methods for PDE's. Brush up heavily on linear algebra, series, etc... Remember the N-body problem is not your friend
(if you solve it give me a call ). So there are things like the Hartree-Fock,
http://en.wikipedia.org/wiki/Hartree%E2%80%93Fock_method
http://en.wikipedia.org/wiki/Post-Hartree%E2%80%93Fock
http://en.wikipedia.org/wiki/Unrestricted_Hartree%E2%80%93Fo...
But oh which theory to use, here's one that people have been using lately to my understanding,
http://en.wikipedia.org/wiki/Density_functional_theory
Of course there are literally hundreds if not thousands of techniques to implement DFT in different ways into soft-ware...
You'll also need your thermodynamics down. I don't mean ΔG = ΔH - T*ΔS. Useful predictions can be performed on an electronic basis, but
turning that into thermodynamics is very important... Edit - Oops just saw you were interested in catalysts. Definitely going to want to be firm on
kinetics! Turning the PDE's into ODE's etc. Steady state approximations. Hmmm, could see an algorithm for that, would be an effort though. I'm sure
people have already developed clever ways around a lot of this.
Might be easier to start with some open-source ab initio chemistry library or something similar. Then again if you don't know what the algorithms are
doing, it could be a whole lot harder to implement them.
If you wish to implement visualization in 3 space, you'll have to decide on DirectX or OpenGl. I prefer OpenGl as it's cross plat-form. GLSL is also
just slightly more intuitive to me personally than HLSL. I'm sure many people would scoff at that, and say Direct X. From there, either write you're
own graphics engine or again find an open source one, hopefully tailored for speed and modify what you need.
For the processing of calculations I think a lot of people are actually doing this on GPU's now days rather then the CPU. http://en.wikipedia.org/wiki/Molecular_modeling_on_GPUs
Hope this gives you some ideas.
[Edited on 25-5-2014 by smaerd]
[Edited on 26-5-2014 by smaerd]
|
|
The Volatile Chemist
International Hazard
Posts: 1981
Registered: 22-3-2014
Location: 'Stil' in the lab...
Member Is Offline
Mood: Copious
|
|
That indeed was what I needed! It'll get me started this summer (School's almost out for me, then I'll have much more free time!).
As I side note, I'm working on a scientific 'calculator'/'assistant' console program. It's tailored to my needs, but when I'm finished it may find use
for others.
|
|
quantumchromodynamics
Hazard to Self
Posts: 67
Registered: 25-9-2013
Location: with much determination, nowhere in particluar
Member Is Offline
Mood: tired but still trying
|
|
amazing super chemistry software
At the recommendation of a friend I have been learning about,
http://www.chemaxon.com/
MarvinSketch.
I have learned more o-chem using this program first, then looking up what I don't understand in my books and online. If one is interested in chemistry
software, this program is quite interesting.
|
|
smaerd
International Hazard
Posts: 1262
Registered: 23-1-2010
Member Is Offline
Mood: hmm...
|
|
Marvin sketch is great for laying out reaction schemes, quick modeling, theoretical calculations, etc. It does have a nice resonance structure creator
which I have used for quickly checking substrates for dominant resonance forms for possible reactivity. Not really sure how it could teach organic
chemistry though. It is my preferred soft-ware suite though, so I feel you there.
|
|
yiberkit
Harmless
Posts: 25
Registered: 13-7-2014
Member Is Offline
Mood: No Mood
|
|
Quote: Originally posted by ampakine | I'm studying chemistry in college but I seem to have more of a passion for programming and computer related things so I'm gaining just as much
programming and computer skills as I am chemistry skills these days. I want to get into a chemistry/pharmacology related field but I'm wondering if my
programming skills will come in useful. I'm getting lots of ideas for ways for developing chemistry software and technology and I don't just mean
software for making models such as ChemDraw, I mean developing hardware and software for automating chemistry processes. One example off the top of my
head: I could develop a machine that will automatically determine the solubility of a compound in a wide range of different compounds at different
temperatures. If I'm not mistaken this kinda thing is usually done manually by a chemist. With the right software a machine could do it far more
effectively than a human.
Any other programmers/technophiliacs here? |
Can you write a mod/program and create system for doing routine quality control analysis in the chemical labs for industry?
If you do , it would be much more helpful for industrialists...
|
|
Pages:
1
2 |