Re: [DIYbio] Re: Medicine Simulation Newbie

If you are looking for an interesting protein modeler, here is the winner of iGEM undergrad this year. Amazing project that circularizes proteins, provides a tool to allow you to design and do it yourself and provides lots of computing power to do so. 




On Nov 27, 2014, at 2:35 PM, Josiah Zayner <josiah.zayner@gmail.com> wrote:

Ujjwal,

     I think you are looking at this the wrong way. It is not possible in most Scientific endeavours to know all possible outcomes. The key is to ask the correct question. If my question is: "Could protein X bind molecule Y?". I can come upon a very reasonable conclusion using Molecular Dynamics(MD).

If my questions is "How does protein X work?" That is also something that is quite possible to uncover and volumes of work have been published on understanding how proteins function through use of Molecular Dynamics(MD) which then were verified later by physical experimentation.

I think MD is a very worthwhile endeavour for someone with programming and computational skills to contribute to Science. I do simulations myself.


On Thu, Nov 27, 2014 at 12:22 PM, Ujjwal Thaakar <ujjwalthaakar@gmail.com> wrote:
Hi Josiah,
Mine was a really abstract question. I basically wanted to get a feel of how productive it is right now to design molecular machinery with preciseness of behaviour through computational biology. I understand I may not be making complete sense to you due to my lack of understanding of the material involved at this point of time. Believe me you're not making much sense to me either :p
Thanks for the pointers though. I do get a feel that while it can be done for small sample sizes and certain conditions, it's yet not possible or even computationally feasible to model out a protein's structural folding and interactions for every possible configuration.

On Fri Nov 28 2014 at 01:42:04 Josiah Zayner <josiah.zayner@gmail.com> wrote:
I think your question is a complicated one. What do you mean by accuracy?

People have "accurately" folded small proteins using simulations as measured by the Root Mean Square Deviation(RMSD) between the backbone(C alpha I think) of the two structures. But still, what do you mean by accurate? A 1 angstrom RMSD? A 0.5 angstrom RMSD? &c. How long.often does the simulation need to sample that state before you view it as significant? 5 picoseconds? 100 picoseconds?

Protein function is a thermodynamic problem. There really is no measure of "accuracy". Proteins are constantly being altered by their environment(i.e. temperature and solvent) see work by Boltzmann and Gibbs to start. What this means is that there are a very high number of possible conformations a protein can undertake. We call this the ensemble. Theoretically an ensemble contains every possible conformation of the protein. Most physical measurements we do of proteins are ensemble measurements. A Molecular Dynamics simulation is a single molecule equilibrium simulation which can't possibly sample all known states of the protein. For a single small protein with 100 amino acids and 3 degrees of freedom for each backbone phi psi angle you are talking 3^198 (This doesn't include sidechains and is probably a gross underestimate) see http://en.wikipedia.org/wiki/Levinthal%27s_paradox. There are techniques to increase the sampling frequency of different states but 3^198 is an awful lot. Your CPU clock frequency is probably 3^9 / second which means is would take about 9^86 years just to count that high without doing any other calculations.

I think the question you might be trying to ask is "Can simulations predict protein function or binding to other molecules?" The answer to that is "Sometimes". It really depends on the thermodynamics of the protein in question and the difficulty in sampling the state you are attempting to understand.

I don't know what you mean by "different environments". Different solvents? With other molecules? In a membrane? All of the above?



Josiah Zayner


On Thu, Nov 27, 2014 at 11:47 AM, Ujjwal Thaakar <ujjwalthaakar@gmail.com> wrote:
To what level of simulation accuracy has computational biology reached? To what degree of accuracy can we predict the behaviour of molecules in different environments?


On Wednesday, November 26, 2014 4:56:47 AM UTC+5:30, Josiah Zayner wrote:
I think first you need to understand some basic biochemistry before you begin.
Unless you are trying to simulate disease spread. Numerous people have worked on this and articles and methods should be easy to find.

If you are trying to understand how drugs help prevent or cure diseases such as Ebola you need to look at the proteins. Here is a tutorial for that: https://docs.google.com/document/d/13H-d8TzDn3m1H2BE_RA617H7g7dBjz_yibCLhLDejSQ/
which can be followed up with simulating the proteins which can be achieved with a tutorial like this: http://web.mit.edu/vmd_v1.9.1/namd-tutorial-unix.pdf


On Tuesday, November 25, 2014 5:58:38 AM UTC-8, Gokula Krishna wrote:
Hello There,

I am a Computer Science Engineer who is interested in simulating vaccines for various disease. Can anyone help me out on how to proceed? I have searched for various journals regarding this, but it is subtle to grasp and they are too abstract. 

This is the basic thesis that I have regarding the program:
1-> The user should specify the disease the various properties about the disease and environment.
2-> The program tests whether the given disease is curable for the vaccine.
3-> What are the parameters should I consider?

Can you suggest me a good place to start with?

Thanks in advance.

--
-- You received this message because you are subscribed to the Google Groups DIYbio group. To post to this group, send email to diybio@googlegroups.com. To unsubscribe from this group, send email to diybio+unsubscribe@googlegroups.com. For more options, visit this group at https://groups.google.com/d/forum/diybio?hl=en
Learn more at www.diybio.org
---
You received this message because you are subscribed to a topic in the Google Groups "DIYbio" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/diybio/QVH2Z3Enu_8/unsubscribe.
To unsubscribe from this group and all its topics, send an email to diybio+unsubscribe@googlegroups.com.
To post to this group, send email to diybio@googlegroups.com.
Visit this group at http://groups.google.com/group/diybio.
To view this discussion on the web visit https://groups.google.com/d/msgid/diybio/40a56e29-fef6-44a1-9b74-50d4bbca7e6d%40googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

--
-- You received this message because you are subscribed to the Google Groups DIYbio group. To post to this group, send email to diybio@googlegroups.com. To unsubscribe from this group, send email to diybio+unsubscribe@googlegroups.com. For more options, visit this group at https://groups.google.com/d/forum/diybio?hl=en
Learn more at www.diybio.org
---
You received this message because you are subscribed to a topic in the Google Groups "DIYbio" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/diybio/QVH2Z3Enu_8/unsubscribe.
To unsubscribe from this group and all its topics, send an email to diybio+unsubscribe@googlegroups.com.
To post to this group, send email to diybio@googlegroups.com.
Visit this group at http://groups.google.com/group/diybio.
To view this discussion on the web visit https://groups.google.com/d/msgid/diybio/CAEUkM4vrUcnYeLJpHbB8dGOTtsS86iT1XpWf2JGxjwk3_ONe5w%40mail.gmail.com.
For more options, visit https://groups.google.com/d/optout.

--
-- You received this message because you are subscribed to the Google Groups DIYbio group. To post to this group, send email to diybio@googlegroups.com. To unsubscribe from this group, send email to diybio+unsubscribe@googlegroups.com. For more options, visit this group at https://groups.google.com/d/forum/diybio?hl=en
Learn more at www.diybio.org
---
You received this message because you are subscribed to a topic in the Google Groups "DIYbio" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/diybio/QVH2Z3Enu_8/unsubscribe.
To unsubscribe from this group and all its topics, send an email to diybio+unsubscribe@googlegroups.com.
To post to this group, send email to diybio@googlegroups.com.
Visit this group at http://groups.google.com/group/diybio.
To view this discussion on the web visit https://groups.google.com/d/msgid/diybio/CAJ1yqBZCbYPen4TPnXYJdRxCj3Jo9_hSvgwEMJ0EAiV7%2Bz13eg%40mail.gmail.com.

For more options, visit https://groups.google.com/d/optout.

--
-- You received this message because you are subscribed to the Google Groups DIYbio group. To post to this group, send email to diybio@googlegroups.com. To unsubscribe from this group, send email to diybio+unsubscribe@googlegroups.com. For more options, visit this group at https://groups.google.com/d/forum/diybio?hl=en
Learn more at www.diybio.org
---
You received this message because you are subscribed to the Google Groups "DIYbio" group.
To unsubscribe from this group and stop receiving emails from it, send an email to diybio+unsubscribe@googlegroups.com.
To post to this group, send email to diybio@googlegroups.com.
Visit this group at http://groups.google.com/group/diybio.
To view this discussion on the web visit https://groups.google.com/d/msgid/diybio/CAEUkM4s-fgcUSwHFhPLnU-q_b58oec53d5bVbY2U8rmAmDrjWg%40mail.gmail.com.
For more options, visit https://groups.google.com/d/optout.

  • Digg
  • Del.icio.us
  • StumbleUpon
  • Reddit
  • RSS

0 comments:

Post a Comment