Is it too late to start new one?

[kyhekm]

Hi

So I got in real big trouble now. I was gonna research about zebrafish so I contacted a grad student in near college.
At first, I thought he said he would help me but he just told me (a week ago) that he doesn't have time and lab space to help me.

so I am kinda in panic right now because I've done lots of research and doesn't have any space.

Do you think I can change the topic now and make it before March?
I guess I'm having the worst Christmas present I've ever got... haha


P.S. Merry Christmas everyone!!

[deleted-71417]

Hi,

You still have 2 ½ months to the science fair. That is plenty of time to do many science fair projects.
Check out the Science Buddies Project Idea guide for lots of interesting options:

https://www.sciencebuddies.org/science- ... deas.shtml

Your other option is to try to find a suitable mentor that can help you.. Check out this advice on finding a mentor:

https://www.sciencebuddies.org/science- ... tors.shtml

Between these two approaches you still have time to do a great project. I wisj you the best of success! And do not forget to have fun doing it.

Best regards,


Barrett L Tomlinson

[aelin]

Hi,

Just another small bit of advice. So if you still want to continue doing the same research, as barrettomlinson suggested, you have to find a new mentor; those guidelines are rather helpful.

However, if you do wish to switch projects for the upcoming fair (and then do the zebrafish research next year maybe), consider various ways to downsize either your project or another project that you are interested. A few years back, I ran into a similar problem as you; I had designed an experiment after gathering a ton of info but then realized that there was a fatal flaw in it when I talked to a prospective mentor. So, with just a few months left, I constructed a quick but well-designed computer model of the experiment via clinical data and some of my own equations, and I did quite well at the fair.

I actually think that computer models are a pretty good option with just a few months left for a few reasons:
1) Time- you can work on your model anywhere and everywhere, so you have an unlimited amount of time to work on it, whereas an experiment is limited by the access to a lab.
2) Data- you can get clinical data and fit it to your model, which precludes potential setbacks in experimentation; this gets you more data in a shorter period of time
3) Presentability- this is a debatable point, but I found it relatively simple to present my model because I was able to draw out a clear experimental method and had pictures of the model running

There is, of course, the difficulty of dealing with models though. There is a quote that goes something like, "Models are never accurate, but sometimes useful," so you have to be ready to defend both the validity and applicability of your model. Validity can be demonstrated with sample clinical data sets, but you must be careful about how you word your conclusions, especially when models are involved. You also have to consider factors that your model did not account for (it is sometimes to exclude certain factors so long as you have a reasonable explanation as to how it will affect the results; this depends on the importance of said factors though), so you have to be very careful with presenting models.

Hope this helps!
Aaron Lin

[kyhekm]

Hi
Thanks for replying to my post, barretttomlinson and aelin.

I was waiting for a professor's e-mail till today because she was out of office till 1/08. I thought she would reply in this week but I guess she just igonored
my e-mail haha..
So I'm thinking about starting new project.
I have a question to you, aelin. I think computer model sounds good for me right now
because I don't have much time but I can do computer model project whenever I want to.
Thanks for suggesting woderful idea, aelin.

aelin, I actually don't know how to make computer model.
For example, if I want to make a model for regeneration of zebrafish, where can I get the clinical datas and how do I make computer model?
I am good at using computer like other teenagers... but I don't know how to make a computer model. So can you give me an example? or what you did?

and when you make a computer model, do I need to make my own creative equation? and are there any programs what you used for your project, aelin?

Thank you.

[aelin]

Hi,

Hm yeah, that's unfortunate about the lab. It happens sometimes though... sometimes they just get too many emails, try re-sending it once, but I suspect that it might be a little late before she responds.

Anyways, about the model.

First, you have to very specifically pick what you want to model/simulate, and this will have to relate to your research question (eg, what you are trying to learn or test). I'm not too familiar with zebrafish regeneration, so I'm just going to make up a scenario. For example, you might want to model the rate at which regeneration occurs on a macroscopic level as a function of biochemical interactions at the microscopic level. Or, you could try to model the survival rate as a function of rate of regeneration. Again, I'm not too familiar with the topic, so you'll have to figure out most of this part on your own. That being said, try and keep it relatively simple, keeping in mind that you will have to implement a method to actually simulate this, so don't go overboard crazy.

Next, you have to decide what type of model to make. You can leave it as a simple mathematical model (see the one here for an example: http://www.uni-tuebingen.de/modeling/Prog/SIR.html), or you can go for a full-on visual simulation. Both have merits, and I find that some combination works best in terms of being able to present the model/simulation. This will in part depend on what you want to model -- if you have too many actors running in the simulation at once, it gets very messy (you'll see what I mean in a minute).

As for the actual modeling. To be honest, I'm very bad with computers, and going into the project, I was just learning how to program (I was taking AP Comp Sci at the same time). But, the complexity of the model will mostly come from the actual bio knowledge that you put in, not the programming (you could have very pretty looking code, but the judges won't really care, they'll all be looking at the bio part). So my code was pretty crappy, but the program worked, and I got data out, which was really all that was important. If you can't already program (I used Java, but C++ or another language would be just as good), talk to your comp sci teacher to see if he can give you a crash course in programming. I did the vast majority of the programming myself but used GridWorld (http://www.horstmann.com/gridworld/) to create a slightly nicer looking display (I used grids, so this was also rather convenient). But like I said, depending on what you want to do, the display and implementation will change. But from the display and basic grid functions, you can just kind of insert the formulas and equations where it seems to make sense that they are executed. As a result of the way the program runs, you will sometimes need to modify equations or create them in a logical way so that the model/simulation proceeds like an actual experiment would.

So this is what I did. I was interested in modeling superinfection (multiple infection of a single cell) dynamics on a multicellular scale. In short, my model had roughly 300 susceptible cells in the initial condition, as well as a few infected ones, and had the viruses replicate. In the experimental trials, superinfection was allowed to occur within some of the cells, but in the control trials, superinfection was not allowed. With each successive time step, cells and viruses would move around and interact as if they were in a petri dish. At the end of each trial, I examined how many strains of the virus there were and how virulent each one was to determine what the overall trend was as a result of superinfection dynamics. In terms of the modeling, I had to modify some of the equations that dealt with replication rate, etc since time steps aren't a particularly quantitative sort of measure. There were some equations in the literature for competitive superinfection dynamics within a single cell, so I utilized those at the single cell level (but those didn't exist at the multicellular level, which was why I was modeling the multicellular scale). The part where I had to invent some equations was later on in the cell-to-cell transmission of virus since I couldn't find any literature there, so I ended up using a modified version of the regular transmission equation (it made sense at the time since the biochemical routes of transmission are similar, but it is easier via cell-to-cell transmission). It's a little big vague from the description, but I can send the model to you as a Java file if you would like.

Going back to something earlier, I have a feeling that the biochemical pathway modeling up to macroscale might be most useful (but a bit more difficult) in your case. It would be somewhat similar to what I was doing in that you could have the chemical intermediates and enzymes function as agents moving throughout space, and they can interact in a probabilistic fashion based on particle collision equations and whatnot. Ultimately, all of those intermediates would form the final product that causes cellular and tissue regeneration, or something like that. Again, though, it is your choice and your project of what you choose to model.

Oh, and about the clinical data. So I validated my model by testing it with clinical data. So for my project, I pulled data from case studies of patients with superinfection by different strains of viruses (this actually is not that uncommon) and input some of the 'initial' data for the patients into the model, pretending that a blood sample or something had been drawn. I then ran that initial data though the model and compiled averages of the results to see if it matched up with the final condition of the patients. This part was actually the trickiest since sometimes, patients just have strange things happen to them, and sometimes, the model is just way wrong, so you have to keep tinkering. You could probably find regeneration data regarding zebrafish in various papers that are floating around, though again, it is tricky to find.

Feel free to ask more questions though, but I think you're on the right track.
Hope this helps!
Aaron Lin