/-/https/www.sciencebuddies.org/cdn/Files/21464/5/sb-logo.png){kind=logo}
/-/https/www.sciencebuddies.org/cdn/Files/21464/5/sb-logo.png){kind=logo}
Hello,
My project is on modeling or possibly creating an algorithm to show the different biochemical pathways that epilepsy evokes to cause neuronal cell death.
If I do a model, it seems I should use stochastic techniques, as both the brain and epilepsy are very dynamic environments. However, none of the articles I have read show which programs they used. Or if they did, the programs aren't free.
So I think a better bet would be to write an algorithm. That way I don't have to study extra calculus... I would have to acquire more programming skills, but there is so much information out there that I think that'll be OK. But what programming language would be the most conducive for my goal? BioPerl is the most common, followed by Python and C. I know a bit of C...
I'm not quite sure where to start either. Is this project even feasible?
Bioinformatics help -- what software would be helpful?
Moderators: AmyCowen, kgudger, MadelineB, Moderators
-
deleted-71949
- Posts: 59
- Joined: Mon Jun 22, 2009 3:55 pm
- Occupation: Student: High School Senior
- Project Question: Applying evolutionary algorithms to a Monte Carlo Tree Search
- Project Due Date: September 20, 2013 (to allow for snail mail, etc.)
- Project Status: I am conducting my experiment
-
deleted-71882
- Former Expert
- Posts: 338
- Joined: Sun Jul 10, 2011 5:48 pm
- Occupation: retired physicist
- Project Question: n/a
- Project Due Date: n/a
- Project Status: Not applicable
Re: Bioinformatics help -- what software would be helpful?
Hello blueswim,
The project you outline is certainly an ambitious one. I'm not an expert in biosystem modeling, but I know about modeling in other systems.
To say that you want to "model" cell death leaves much to my imagination. Neurons die because of so many reasons. Large numbers of chemical components are involved, and their presence in the cell is due to many biochemical pathways extending throughout the organism.
To get an idea of what you have in mind, I did a quick web search and turned up http://www.biotech.iitm.ac.in/faculty/d ... _death.pdf.
Please look at this paper and tell me if the approaches in the paper are at all like what you have in mind. If not, please give me another reference or some specifics of what variables and events you would use in your model.
Yes, the project is feasible if you know enough about neuronal cell biology and can spend enough time on the project.
Don't worry about programming the model until you have figured out what you want to accomplish. Any of a dozen languages will do just fine, and lots of information about programming with any of them is available on the web. However, be warned that programming can devour large amounts of your time.
Good luck. Keep in touch, WW
The project you outline is certainly an ambitious one. I'm not an expert in biosystem modeling, but I know about modeling in other systems.
To say that you want to "model" cell death leaves much to my imagination. Neurons die because of so many reasons. Large numbers of chemical components are involved, and their presence in the cell is due to many biochemical pathways extending throughout the organism.
To get an idea of what you have in mind, I did a quick web search and turned up http://www.biotech.iitm.ac.in/faculty/d ... _death.pdf.
Please look at this paper and tell me if the approaches in the paper are at all like what you have in mind. If not, please give me another reference or some specifics of what variables and events you would use in your model.
Yes, the project is feasible if you know enough about neuronal cell biology and can spend enough time on the project.
Don't worry about programming the model until you have figured out what you want to accomplish. Any of a dozen languages will do just fine, and lots of information about programming with any of them is available on the web. However, be warned that programming can devour large amounts of your time.
Good luck. Keep in touch, WW
-
deleted-71949
- Posts: 59
- Joined: Mon Jun 22, 2009 3:55 pm
- Occupation: Student: High School Senior
- Project Question: Applying evolutionary algorithms to a Monte Carlo Tree Search
- Project Due Date: September 20, 2013 (to allow for snail mail, etc.)
- Project Status: I am conducting my experiment
Re: Bioinformatics help -- what software would be helpful?
Hello,
I forgot to mention that the neuronal death I want to investigate is apoptosis. Well, actually I changed my topic a little bit. Epilepsy is a broad disease, so I narrowed it down to traumatic brain injury (TBI) induced epilepsy of the temporal lobe. TBI induced epilepsy goes unnoticed sometimes, because patients believe their injury has long recovered, as well as the fact that the seizures are typically what are known as "simple seizures" - it doesn't cause loss of consciousness. Epilepsy of the temporal lobe, especially is hard to treat - only 30% respond to medication. Repetitive seizures can actually cause neuronal death, so there's a need to protect neurons. Seizure induced neuronal death is highly evidenced to be caused by pro-apoptotic signals being induced. TBI induced brain death is necrosis.
People don't know why the seizures cause neuronal death, so I want to try to see how TBI induced epilepsy (at least the signaling pathways causing it) interact with the apoptotic and/or ion channel pathways.
The article isn't really what I had in mind, though it seems interesting. The article I am now planning to reference for procedures is here: http://www.biomedcentral.com/1471-2105/10/370
I think that the article is discussing interactions between two apoptotic pathways, but I am not certain. However, I believe the approach can be used for my project.
Thanks,
blueswim
I forgot to mention that the neuronal death I want to investigate is apoptosis. Well, actually I changed my topic a little bit. Epilepsy is a broad disease, so I narrowed it down to traumatic brain injury (TBI) induced epilepsy of the temporal lobe. TBI induced epilepsy goes unnoticed sometimes, because patients believe their injury has long recovered, as well as the fact that the seizures are typically what are known as "simple seizures" - it doesn't cause loss of consciousness. Epilepsy of the temporal lobe, especially is hard to treat - only 30% respond to medication. Repetitive seizures can actually cause neuronal death, so there's a need to protect neurons. Seizure induced neuronal death is highly evidenced to be caused by pro-apoptotic signals being induced. TBI induced brain death is necrosis.
People don't know why the seizures cause neuronal death, so I want to try to see how TBI induced epilepsy (at least the signaling pathways causing it) interact with the apoptotic and/or ion channel pathways.
The article isn't really what I had in mind, though it seems interesting. The article I am now planning to reference for procedures is here: http://www.biomedcentral.com/1471-2105/10/370
I think that the article is discussing interactions between two apoptotic pathways, but I am not certain. However, I believe the approach can be used for my project.
Thanks,
blueswim
-
deleted-71882
- Former Expert
- Posts: 338
- Joined: Sun Jul 10, 2011 5:48 pm
- Occupation: retired physicist
- Project Question: n/a
- Project Due Date: n/a
- Project Status: Not applicable
Re: Bioinformatics help -- what software would be helpful?
blueswim,
Thanks for the reference. Now I have a much better idea about the sort of thing you want to do.
I don't think the reference paper uses any math beyond what you get in a good high-school course. On the other hand, the work uses lots of math, so you need to be good at figuring out what's going on, and you need to be good at applying what you know and learning new, slightly different versions of what you already know. I think all the algorithms and numerical methods will be discussed well on the web.
I emphasize that the project will require a lot of time. Reproducing the work in the reference for your system-of-interest sounds more like a suitable topic for a graduate thesis than a typical science project. Your project does sound like the sort of thing that wins in big-name competitions.
Consider carefully whether you can devote so much time to your project while meeting all your other commitments. You might set a time limit to work on this project with the options at the end of that time to either commit to finishing this project or bailing out and choosing a much simpler, quicker project to satisfy any assigned commitments.
I note that the authors of the reference paper say, "Source code available on request." I suggest that you e-mail one of them and request the source code. Look over the code and determine whether you can understand it or not. I gather you haven't done such a task before. I can give you some tips on how to get started. It's a strange sort of task to understand somebody else's code. It's much easier if the author has inserted understandable comments. It's nearly impossible if no comments were used.
I hope and strongly suspect that their software will run on a modern laptop.
If you can make some headway and understand the reference paper and some of the code, then you can begin to make a list of the tasks you have to perform before you will have a working version of the code that is tailored to your problem and loaded with data for your problem.
Let me know what you plan to do, and what progress you make. I will try to offer help on understanding the paper and the code.
Again, good luck and applause for your high goal.
WW
Thanks for the reference. Now I have a much better idea about the sort of thing you want to do.
I don't think the reference paper uses any math beyond what you get in a good high-school course. On the other hand, the work uses lots of math, so you need to be good at figuring out what's going on, and you need to be good at applying what you know and learning new, slightly different versions of what you already know. I think all the algorithms and numerical methods will be discussed well on the web.
I emphasize that the project will require a lot of time. Reproducing the work in the reference for your system-of-interest sounds more like a suitable topic for a graduate thesis than a typical science project. Your project does sound like the sort of thing that wins in big-name competitions.
Consider carefully whether you can devote so much time to your project while meeting all your other commitments. You might set a time limit to work on this project with the options at the end of that time to either commit to finishing this project or bailing out and choosing a much simpler, quicker project to satisfy any assigned commitments.
I note that the authors of the reference paper say, "Source code available on request." I suggest that you e-mail one of them and request the source code. Look over the code and determine whether you can understand it or not. I gather you haven't done such a task before. I can give you some tips on how to get started. It's a strange sort of task to understand somebody else's code. It's much easier if the author has inserted understandable comments. It's nearly impossible if no comments were used.
I hope and strongly suspect that their software will run on a modern laptop.
If you can make some headway and understand the reference paper and some of the code, then you can begin to make a list of the tasks you have to perform before you will have a working version of the code that is tailored to your problem and loaded with data for your problem.
Let me know what you plan to do, and what progress you make. I will try to offer help on understanding the paper and the code.
Again, good luck and applause for your high goal.
WW
-
deleted-71949
- Posts: 59
- Joined: Mon Jun 22, 2009 3:55 pm
- Occupation: Student: High School Senior
- Project Question: Applying evolutionary algorithms to a Monte Carlo Tree Search
- Project Due Date: September 20, 2013 (to allow for snail mail, etc.)
- Project Status: I am conducting my experiment
Re: Bioinformatics help -- what software would be helpful?
Thank you for the reply. I have sent out an email to author #1 because the article notes that she had the idea for the project.
As I read through the paper, it seems there are four major steps:
As I read through the paper, it seems there are four major steps:
- 1) Database search for proteins of interest and signaling pathways
2) Data mining - this can be done in Python/C or specialized software such as Orange (but still uses both Python and C)
3) Calculations of pathways
4) Connections between different pathways
(Steps 3 and 4 seemed to have been using software that analyzes networks (not just in biological ones)
-
deleted-71882
- Former Expert
- Posts: 338
- Joined: Sun Jul 10, 2011 5:48 pm
- Occupation: retired physicist
- Project Question: n/a
- Project Due Date: n/a
- Project Status: Not applicable
Re: Bioinformatics help -- what software would be helpful?
blueswim,
Fast work!
Let's call your reference paper INP from the authors last-name initials to save keystrokes INP use several software packages to support their code. It looks like most or all of them are available for free. UCINET might cost you $40. If not free or cheap, perhaps some other package can substitute.
Is your goal to just describe the pathways and list the connectivity, etc. as they do, or do you have in mind to use the network to somehow simulate TBI as activation and/or inhibition of certain signals and compute the net effect on a target neuron? Or something else?
WW
Fast work!
Let's call your reference paper INP from the authors last-name initials to save keystrokes INP use several software packages to support their code. It looks like most or all of them are available for free. UCINET might cost you $40. If not free or cheap, perhaps some other package can substitute.
Is your goal to just describe the pathways and list the connectivity, etc. as they do, or do you have in mind to use the network to somehow simulate TBI as activation and/or inhibition of certain signals and compute the net effect on a target neuron? Or something else?
WW
-
deleted-71949
- Posts: 59
- Joined: Mon Jun 22, 2009 3:55 pm
- Occupation: Student: High School Senior
- Project Question: Applying evolutionary algorithms to a Monte Carlo Tree Search
- Project Due Date: September 20, 2013 (to allow for snail mail, etc.)
- Project Status: I am conducting my experiment
Re: Bioinformatics help -- what software would be helpful?
All right! Sounds like a plan.
My goal is to see how combinations of different activation/inhibition signals in the network determine what the fate of a neuron is (i.e. whether it dies, and if it dies, the mechanism of death - apoptosis or necrosis). This will allow for the discovery of new targets for protecting the neurons (i.e. when the fate is death). The main, first goal, however, is to establish that there is a clear link between neuronal death and the epileptic seizures caused by TBI.
blueswim
My goal is to see how combinations of different activation/inhibition signals in the network determine what the fate of a neuron is (i.e. whether it dies, and if it dies, the mechanism of death - apoptosis or necrosis). This will allow for the discovery of new targets for protecting the neurons (i.e. when the fate is death). The main, first goal, however, is to establish that there is a clear link between neuronal death and the epileptic seizures caused by TBI.
blueswim
Last edited by deleted-71949 on Tue Aug 07, 2012 10:15 pm, edited 1 time in total.
-
deleted-71949
- Posts: 59
- Joined: Mon Jun 22, 2009 3:55 pm
- Occupation: Student: High School Senior
- Project Question: Applying evolutionary algorithms to a Monte Carlo Tree Search
- Project Due Date: September 20, 2013 (to allow for snail mail, etc.)
- Project Status: I am conducting my experiment
Re: Bioinformatics help -- what software would be helpful?
Hello,
I haven't got a reply from the author yet, so I decided to do some more research.
I found these articles:
And then I found these:
They (or at least the third one, which I read almost to the end) seem to use BioPython, a subset of Python.
blueswim
I haven't got a reply from the author yet, so I decided to do some more research.
I found these articles:
- http://ccforum.com/content/9/1/66
http://www.meb.uni-bonn.de/agBeck/cms/u ... Review.pdf
http://www.biolcell.org/boc/095/0329/boc0950329.pdf
And then I found these:
- http://repository.ias.ac.in/4390/1/327.pdf
http://www.comp.nus.edu/~wongls/psZ/apb ... bc101a.pdf
http://www.ipcbee.net/vol5/58-X00091.pdf
They (or at least the third one, which I read almost to the end) seem to use BioPython, a subset of Python.
blueswim
-
deleted-71882
- Former Expert
- Posts: 338
- Joined: Sun Jul 10, 2011 5:48 pm
- Occupation: retired physicist
- Project Question: n/a
- Project Due Date: n/a
- Project Status: Not applicable
Re: Bioinformatics help -- what software would be helpful?
blueswim,
If you don't hear from the lead author, try the last one. The last author is usually the most senior one. Since you don't have a research institution affiliation, be sure to tell them why you want the software and a bit about your project. Most researchers are willing to help students.
In any case, you don't have to wait on a reply. I noticed that the paper contains a link to the COSBI software in the Methods section. That link is bad, but a little detective work turned up the correct one, http://www.cosbi.eu/index.php/research/prototypes/graph. A user manual and a download link are provided there.
Good research. You want to be sure that you understand how your project relates to previous work, and you want to know that your work adds to what is already known. If you were doing a Ph.D. thesis, it would be essential to not replicate previous work. Ph.D. reasearch must be original. I don't know what originality requirements a science fair might have.
Are you aiming for a local science fair or one of the big ones?
After a quick glance, the references seem to describe variations on the modeling scheme, but they all seem to follow the same scheme.
Get data,
Build network,
Look at what the network does.
INF seems to do very little of the last step. It seems to me that the edges of the network graph need to determine specifically how the connection works. Then you could specify a time-variant input of the signaling species and compute the time-variant output concentration of signal(s) to a target neuron. Do you see this step implemented in any of the references?
While I understand the ideas in INF, I can't judge how your project will fit into the overall current research on seizure-induced cell death or the wider issue of post-TBI treatment options. Do you have an advisor that can help with these issues? Do you live near a good medical school or university where you might find such an advisor?
Keep going on the research.
WW
If you don't hear from the lead author, try the last one. The last author is usually the most senior one. Since you don't have a research institution affiliation, be sure to tell them why you want the software and a bit about your project. Most researchers are willing to help students.
In any case, you don't have to wait on a reply. I noticed that the paper contains a link to the COSBI software in the Methods section. That link is bad, but a little detective work turned up the correct one, http://www.cosbi.eu/index.php/research/prototypes/graph. A user manual and a download link are provided there.
Good research. You want to be sure that you understand how your project relates to previous work, and you want to know that your work adds to what is already known. If you were doing a Ph.D. thesis, it would be essential to not replicate previous work. Ph.D. reasearch must be original. I don't know what originality requirements a science fair might have.
Are you aiming for a local science fair or one of the big ones?
After a quick glance, the references seem to describe variations on the modeling scheme, but they all seem to follow the same scheme.
Get data,
Build network,
Look at what the network does.
INF seems to do very little of the last step. It seems to me that the edges of the network graph need to determine specifically how the connection works. Then you could specify a time-variant input of the signaling species and compute the time-variant output concentration of signal(s) to a target neuron. Do you see this step implemented in any of the references?
While I understand the ideas in INF, I can't judge how your project will fit into the overall current research on seizure-induced cell death or the wider issue of post-TBI treatment options. Do you have an advisor that can help with these issues? Do you live near a good medical school or university where you might find such an advisor?
Keep going on the research.
WW
-
deleted-71949
- Posts: 59
- Joined: Mon Jun 22, 2009 3:55 pm
- Occupation: Student: High School Senior
- Project Question: Applying evolutionary algorithms to a Monte Carlo Tree Search
- Project Due Date: September 20, 2013 (to allow for snail mail, etc.)
- Project Status: I am conducting my experiment
Re: Bioinformatics help -- what software would be helpful?
Thanks!
I am starting at the local science fairs, i.e. county, BioGENEious. But I also plan to submit this project to such things as JSHS. I wanted to submit to Siemens, but I need a partner. At the local science fairs, I want to win some awards, and perhaps go on to the next level. Next summer, hopefully I can go to a research program.
Actually, I realized that this article, http://www.ipcbee.net/vol5/58-X00091.pdf, is authored by INF and it is a more in-depth version of the first article.
I don't see the step of time-variant input/outputs in any references I found.
I don't have an advisor as of now. I live within driving distance of UCSF and Stanford. Berkeley may be a bit of a stretch, but might work.
blueswim
I am starting at the local science fairs, i.e. county, BioGENEious. But I also plan to submit this project to such things as JSHS. I wanted to submit to Siemens, but I need a partner. At the local science fairs, I want to win some awards, and perhaps go on to the next level. Next summer, hopefully I can go to a research program.
Actually, I realized that this article, http://www.ipcbee.net/vol5/58-X00091.pdf, is authored by INF and it is a more in-depth version of the first article.
I don't see the step of time-variant input/outputs in any references I found.
I don't have an advisor as of now. I live within driving distance of UCSF and Stanford. Berkeley may be a bit of a stretch, but might work.
blueswim
-
deleted-71882
- Former Expert
- Posts: 338
- Joined: Sun Jul 10, 2011 5:48 pm
- Occupation: retired physicist
- Project Question: n/a
- Project Due Date: n/a
- Project Status: Not applicable
Re: Bioinformatics help -- what software would be helpful?
blueswim,
You couldn't do any better than Stanford or UCSF. You might write to some of the faculty in relevant departments who do research in biochemical signalling or something else related to your problem.
The new INF paper you linked is the written version of an oral presentation given at a conference. At first glance, it seems to discuss only crosstalk in more detail. It is very common for authors to get all the publication mileage out of their work as they can manage. Even presenting the same work in a journal and at a conference is generally acceptable.
I'm puzzled that the models don't seem to specify what happens inside the network. They just give nodes and pathways. What if one pathway responds to an input to send a certain signal to the target, but another pathway immediately sends back a second signal to inhibit the first signal? Doesn't one need to account for this sort of thing?
Since this part of modeling seems so obvious (at least to a novice like me), I think we may be missing something. If the details of what happens along a pathway hasn't been introduced into biochemical signalling models, then there's a great project in just modeling with application to all biochemical signal modeling.
BTW, I wouldn't say that BioPython is a subset of Python. BioPython is a collection of applications for manipulating bioinformation, mainly sequences that is written in Python.
WW
You couldn't do any better than Stanford or UCSF. You might write to some of the faculty in relevant departments who do research in biochemical signalling or something else related to your problem.
The new INF paper you linked is the written version of an oral presentation given at a conference. At first glance, it seems to discuss only crosstalk in more detail. It is very common for authors to get all the publication mileage out of their work as they can manage. Even presenting the same work in a journal and at a conference is generally acceptable.
I'm puzzled that the models don't seem to specify what happens inside the network. They just give nodes and pathways. What if one pathway responds to an input to send a certain signal to the target, but another pathway immediately sends back a second signal to inhibit the first signal? Doesn't one need to account for this sort of thing?
Since this part of modeling seems so obvious (at least to a novice like me), I think we may be missing something. If the details of what happens along a pathway hasn't been introduced into biochemical signalling models, then there's a great project in just modeling with application to all biochemical signal modeling.
BTW, I wouldn't say that BioPython is a subset of Python. BioPython is a collection of applications for manipulating bioinformation, mainly sequences that is written in Python.
WW
-
deleted-71949
- Posts: 59
- Joined: Mon Jun 22, 2009 3:55 pm
- Occupation: Student: High School Senior
- Project Question: Applying evolutionary algorithms to a Monte Carlo Tree Search
- Project Due Date: September 20, 2013 (to allow for snail mail, etc.)
- Project Status: I am conducting my experiment
Re: Bioinformatics help -- what software would be helpful?
OK, good idea.
I think that INF describes what is known as a connection map, but what I want is a simulation - according to http://dp.univr.it/~laudanna/Systems%20 ... tworks.pdf
However, I can't find articles describing anything other than in general.
blueswim
I think that INF describes what is known as a connection map, but what I want is a simulation - according to http://dp.univr.it/~laudanna/Systems%20 ... tworks.pdf
However, I can't find articles describing anything other than in general.
blueswim
-
deleted-71949
- Posts: 59
- Joined: Mon Jun 22, 2009 3:55 pm
- Occupation: Student: High School Senior
- Project Question: Applying evolutionary algorithms to a Monte Carlo Tree Search
- Project Due Date: September 20, 2013 (to allow for snail mail, etc.)
- Project Status: I am conducting my experiment
Re: Bioinformatics help -- what software would be helpful?
Actually, I think what I might really need to be modeling is a signaling cascade. As it is described in http://www.ploscompbiol.org/article/inf ... bi.1000041, it sounds similar to the goal of adding input (i.e. off/on).
I need to factor in the transduction, which is what happens when there is a "talk-back" from the target. I think this model does that: http://www.ncbi.nlm.nih.gov/pmc/articles/PMC1913166/
blueswim
I need to factor in the transduction, which is what happens when there is a "talk-back" from the target. I think this model does that: http://www.ncbi.nlm.nih.gov/pmc/articles/PMC1913166/
blueswim
-
deleted-71949
- Posts: 59
- Joined: Mon Jun 22, 2009 3:55 pm
- Occupation: Student: High School Senior
- Project Question: Applying evolutionary algorithms to a Monte Carlo Tree Search
- Project Due Date: September 20, 2013 (to allow for snail mail, etc.)
- Project Status: I am conducting my experiment
Re: Bioinformatics help -- what software would be helpful?
According to a number of articles
Out of those, the only one I've ever heard of is Markov chains. Which one would be best for this project?
blueswim
- http://people.mbi.ohio-state.edu/baguda ... hapter.pdf
http://www.di.unito.it/~horvath/publica ... HoMa11.pdf
http://www.biomedcentral.com/content/pd ... 11-308.pdf
http://www.biomedcentral.com/1752-0509/3/118/
http://www.dcs.ed.ac.uk/pepa/strongerco ... elling.pdf
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2761888/
http://www.inf.ed.ac.uk/teaching/course ... uction.pdf
Out of those, the only one I've ever heard of is Markov chains. Which one would be best for this project?
blueswim
-
deleted-71949
- Posts: 59
- Joined: Mon Jun 22, 2009 3:55 pm
- Occupation: Student: High School Senior
- Project Question: Applying evolutionary algorithms to a Monte Carlo Tree Search
- Project Due Date: September 20, 2013 (to allow for snail mail, etc.)
- Project Status: I am conducting my experiment
Re: Bioinformatics help -- what software would be helpful?
Also, for people I should contact, what types of people? I mean, for professor there's like joint, emeritus, adjunct, assistant, etc. And then there's the "investigator" title too (at UCSF). Which of these would be best to contact?
Thanks,
blueswim
Thanks,
blueswim
-
deleted-71882
- Former Expert
- Posts: 338
- Joined: Sun Jul 10, 2011 5:48 pm
- Occupation: retired physicist
- Project Question: n/a
- Project Due Date: n/a
- Project Status: Not applicable
Re: Bioinformatics help -- what software would be helpful?
blueswim,
Well, your user name suggests an apt metaphor--you are wading into this topic rather deeply, and I wonder how well can blue swim! Sorry, couldn't resist the silliness.
Seriously, your research is leading you into what is very expectedly a seriously complex issue. So many different physical processes are known to be involved in intercellular signalling, and accounting for each of them involves another piece that has to be added to get a complete model. Right now, I'd say the growing complexity is just an indication that you're doing a good job of research.
Eventually, you will have to confront the circumstance that with your level of knowledge and limited time and resources, you will have to pull back and choose a limited, doable project. Don't rush to that stage, but remember that you have to confront it eventually. Maybe you should set a deadline.
The decision to do this and not that depends crucially on where the overall field of research stands at present. You don't want to choose a project that is old-hat and contains nothing new, and you don't want to go out on a limb so that you don't complete a meaningful result. The best way to get your specific goal is to have the advice of someone who knows the relevant fields and their present status.
So, we're back to the issue of finding an adviser. Maybe you could post to Science Buddies again in Life Science asking for help in finding an adviser. Can't hurt. We've gone on so long in this thread that I suspect no one else is following us.
Direct contact is best, but with whom? Professors come in many flavors:
Joint - has appointments in two or more departments
Emeritus - retired but still puttering around
Adjunct - has a title, but probably not paid and has no chance for advancement. Usually has another job as well.
Assistant - Entry level professor. Can advance to Associate Prof. and then Full Prof.
Investigator - Researcher. Maybe holds other positions as well.
Spend some time looking around the school websites. Perhaps someone who has "professor" in his title would be best, but if you find someone who seems to be working right in the area of interest, contact him for sure. Professors have the most latitude in how they spend their time, but they are also very busy. Maybe you could choose someone in an administrative role in a department of interest and ask them to forward your inquiry to someone who might be able and willing to help you. Indicate that you need assistance in formulating your project, and you do not expect an open-ended commitment to work on the details. Include enough detail about what you already know to indicate that you are a serious and likely successful student researcher. If you can get your foot in the door, then a relationship will bloom or it won't. Make several inquiries: not just one.
Ordinary Differential Equations (ODEs) - Equations that contain not only a variable, but also how fast the variable is changing and accelerating. You need to understand only the most introductory part of calculus (derivatives) to use well-known numerical methods for finding and using the equation solutions.
Kinetic equations - the differential equations of a number of interacting entities. Considerably more complicated than the differential equation of a single entity.
Pi calculus - I had to look up this one http://en.wikipedia.org/wiki/%CE%A0-calculus. Simple, but possibly necessary for a completely realistic signalling network.
No one of these is "best." They are applicable to different situations and are somewhat complementary. I think you should choose the biochemical effects you want to model and then consider the best methods to model them specifically.
WW
Well, your user name suggests an apt metaphor--you are wading into this topic rather deeply, and I wonder how well can blue swim! Sorry, couldn't resist the silliness.
Seriously, your research is leading you into what is very expectedly a seriously complex issue. So many different physical processes are known to be involved in intercellular signalling, and accounting for each of them involves another piece that has to be added to get a complete model. Right now, I'd say the growing complexity is just an indication that you're doing a good job of research.
Eventually, you will have to confront the circumstance that with your level of knowledge and limited time and resources, you will have to pull back and choose a limited, doable project. Don't rush to that stage, but remember that you have to confront it eventually. Maybe you should set a deadline.
The decision to do this and not that depends crucially on where the overall field of research stands at present. You don't want to choose a project that is old-hat and contains nothing new, and you don't want to go out on a limb so that you don't complete a meaningful result. The best way to get your specific goal is to have the advice of someone who knows the relevant fields and their present status.
So, we're back to the issue of finding an adviser. Maybe you could post to Science Buddies again in Life Science asking for help in finding an adviser. Can't hurt. We've gone on so long in this thread that I suspect no one else is following us.
Direct contact is best, but with whom? Professors come in many flavors:
Joint - has appointments in two or more departments
Emeritus - retired but still puttering around
Adjunct - has a title, but probably not paid and has no chance for advancement. Usually has another job as well.
Assistant - Entry level professor. Can advance to Associate Prof. and then Full Prof.
Investigator - Researcher. Maybe holds other positions as well.
Spend some time looking around the school websites. Perhaps someone who has "professor" in his title would be best, but if you find someone who seems to be working right in the area of interest, contact him for sure. Professors have the most latitude in how they spend their time, but they are also very busy. Maybe you could choose someone in an administrative role in a department of interest and ask them to forward your inquiry to someone who might be able and willing to help you. Indicate that you need assistance in formulating your project, and you do not expect an open-ended commitment to work on the details. Include enough detail about what you already know to indicate that you are a serious and likely successful student researcher. If you can get your foot in the door, then a relationship will bloom or it won't. Make several inquiries: not just one.
Ordinary Differential Equations (ODEs) - Equations that contain not only a variable, but also how fast the variable is changing and accelerating. You need to understand only the most introductory part of calculus (derivatives) to use well-known numerical methods for finding and using the equation solutions.
Kinetic equations - the differential equations of a number of interacting entities. Considerably more complicated than the differential equation of a single entity.
Pi calculus - I had to look up this one http://en.wikipedia.org/wiki/%CE%A0-calculus. Simple, but possibly necessary for a completely realistic signalling network.
No one of these is "best." They are applicable to different situations and are somewhat complementary. I think you should choose the biochemical effects you want to model and then consider the best methods to model them specifically.
WW
-
deleted-71949
- Posts: 59
- Joined: Mon Jun 22, 2009 3:55 pm
- Occupation: Student: High School Senior
- Project Question: Applying evolutionary algorithms to a Monte Carlo Tree Search
- Project Due Date: September 20, 2013 (to allow for snail mail, etc.)
- Project Status: I am conducting my experiment
Re: Bioinformatics help -- what software would be helpful?
Hello,
Good news! I already got some replies to my emails for help, and some even answered my questions!
This is exciting!
blueswim
Good news! I already got some replies to my emails for help, and some even answered my questions!
This is exciting!
blueswim
-
deleted-71949
- Posts: 59
- Joined: Mon Jun 22, 2009 3:55 pm
- Occupation: Student: High School Senior
- Project Question: Applying evolutionary algorithms to a Monte Carlo Tree Search
- Project Due Date: September 20, 2013 (to allow for snail mail, etc.)
- Project Status: I am conducting my experiment
Re: Bioinformatics help -- what software would be helpful?
I've now changed my focus - I'm focusing on neuroprotection, because the case I'm talking about happens rarely.
Also, I have decided to focus on spinal muscular atrophy, which I researched a bit earlier this year. The problem is more specific than epilepsy but not so specific that it rarely happens.
blueswim
Also, I have decided to focus on spinal muscular atrophy, which I researched a bit earlier this year. The problem is more specific than epilepsy but not so specific that it rarely happens.
blueswim
-
deleted-71882
- Former Expert
- Posts: 338
- Joined: Sun Jul 10, 2011 5:48 pm
- Occupation: retired physicist
- Project Question: n/a
- Project Due Date: n/a
- Project Status: Not applicable
Re: Bioinformatics help -- what software would be helpful?
blueswim,
I'm pleased to hear that you've made contact with some "real" experts;-)
Let me know if I can help further.
WW
I'm pleased to hear that you've made contact with some "real" experts;-)
Let me know if I can help further.
WW
-
deleted-71949
- Posts: 59
- Joined: Mon Jun 22, 2009 3:55 pm
- Occupation: Student: High School Senior
- Project Question: Applying evolutionary algorithms to a Monte Carlo Tree Search
- Project Due Date: September 20, 2013 (to allow for snail mail, etc.)
- Project Status: I am conducting my experiment
Re: Bioinformatics help -- what software would be helpful?
Hello,
I have changed my project again. I found something that interested me more and seems to be more doable than my other projects.
My topic is now inhibition of multi-drug resistance in a deadly kind of brain cancer, glioblastoma.
I think that my course of action should be:
I have changed my project again. I found something that interested me more and seems to be more doable than my other projects.
My topic is now inhibition of multi-drug resistance in a deadly kind of brain cancer, glioblastoma.
I think that my course of action should be:
- -Find P-gp inhibitors using drug design techniques (as described here: http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2724963/)
-Build a model of glioblastoma tumor cells that will eventually expresses P-gp (using a program called NetLogo, suggested by someone at UCSF)
-Test Avastin IV (gets good relatively good reviews from patients) with and without P-gp inhibitors (also in NetLogo)
-
deleted-71882
- Former Expert
- Posts: 338
- Joined: Sun Jul 10, 2011 5:48 pm
- Occupation: retired physicist
- Project Question: n/a
- Project Due Date: n/a
- Project Status: Not applicable
Re: Bioinformatics help -- what software would be helpful?
blueswim,
Happy to see that you're still learning and working on a project choice. The latest sounds interesting, but I don't see how the two parts--protein structure and then Netlogo modeling fit together--but since this was suggested by UCSF people, it must make sense.
WW
Happy to see that you're still learning and working on a project choice. The latest sounds interesting, but I don't see how the two parts--protein structure and then Netlogo modeling fit together--but since this was suggested by UCSF people, it must make sense.
WW