Forum:Master Thesis topic selection

Question

[hamed2005]

Hi,

I really need your help, for a master thesis (master of Bioinformatics), working on which one do you think is better?

1. Alternative splicing role in cancers
2. Proteogenomics (on a specie with less annotated genome)
3. Drug analysis using machine learning
4. Prediction of protein functions using machine learning
5. Proteomics and Datamining
6. PTM effect on protein interaction (machine learning)

P.S. : I have a bachelor in computer software engineering and I want to start a PhD after finishing my master. I just finished the 1st semester, so I have no deep insight; please help.

[Devon Ryan]

Define "better". They're all fine topics, it's just a matter of what you think is more interesting. I would personally work on 1 or 3, but that's purely personal preference.

[hamed2005]

I'm really interested in all but I'm really looking for something that is really challenging and also helps me for finding a PhD position.

The subjects will be defined in details later since we will start working next year.

[Devon Ryan]

Given your background in CS, the more challenging projects would be those not involving machine learning or any other more CS-centric topics. Since you likely have minimal experience working with actual biological specimens, picking a topic that might allow you to do a couple bench experiments would be nice (analysts are vastly more useful when they have a good feeling for possible sources of experimental noise and bias). Since this is a masters project, obviously don't try to do too much of that though.

Regarding finding a PhD position, you could do that with any of them. Anything cancer related is going to open up more doors than most of the others, since there's always a lot of funding there.

[Michael]

Dear hamed2005. Would be interesting to know how this went. Thesis should be finished by now.

[hamed2005]

I took Devon's advice and selected the last topic in the end, "Effect of PTMs on PPI in bacteria", because it had lab experiments. Well, its not finished yet ;) as I have to do some lab validation and also the work load from courses were really high! But overall I am really happy with my choice!

[WouterDeCoster]

Congratulations, nice work.

Answer 1

[sca.een]

The topic #4 is pretty interesting since predicting the function of an unknown protein is an essential goal in bioinformatics and if you can come up with a new accurate prediction method using computational approaches, you can write an outstanding paper. Actually, the researcher from AceWriters Team recommends this article on identification of protein functions using a machine-learning approach based on sequence-derived properties. Check it out. It is pretty fascinating.