Jumping into bioinformatics

[coiner]

Hello SEQ community!

I recently received a BS in IT and I'm looking to find a place to use my skills. I currently hold a position as a sysadmin for atmospheric research but I really don't think my interests lie there. I know I have a great interest in someday using bioinformatics to predict protein structure and assembly from DNA. I have purchased a few textbooks on the subject of bioinformatics but they always seem so dry and boring to read.

I would like to just dive into some problems as that is the best way that I learn things. Let's say I wanted to start by predicting the protein output of very simple genetic sequences. What would be the best place to start?

[westerman]

I would like to just dive into some problems as that is the best way that I learn things. Let's say I wanted to start by predicting the protein output of very simple genetic sequences. What would be the best place to start?

Since you are coming from the IT side of the world, write a program in your favorite language. Since this task is simple do it de-novo (i.e., without using any packages). Then do it using a package. Test with a variety of genome snippets you can find.

This is different advice than I'd give to a biologist. There I'd say go to Galaxy (or another like tool) and learn how to do protein translation that way.

In both cases, reading this forum backwards and forwards should give you lots of insight into bioinformatics.

And, since it seems like you may be at an institution of higher learning, take some biology/biochemistry classes. I'm not sure if a person can be a bioinformatician without a basic understanding of genetics.

[coiner]

Thanks for the advice. I have taken basic lab sciences involving genetics so I do have some very basic knowledge. I tend to pick things up as I go though so if I could start on a few basic problems and work my way to more complex, I could also learn a lot about biochemistry along the way. I also have many reference texts at my disposal and may perhaps take some classes in the future.

I will certainly read around the forum too! Do you have suggestions for some basic genetic sequences I could work with to get started? and perhaps a common problem that I could solve using those sequences? I'd rather not sift through and get discouraged by diving too deep at the start!

[westerman]

Grab any of the sequences at NCBI. Lots of resources there. Lots of useful tools to read about.

A common problem? Take a set of reads -- look through the SRA (http://www.ncbi.nlm.nih.gov/sra/) and grab a sequencing project, preferably to a well characterized genome. Map the reads to a reference. Figure out what type of contamination the reads -- vector, rRNA. Do a denovo assembly. Try annotating the reads. None of the previous involves programming per se but they is a common task. Try several different programs. There is no one solution.

[GenoMax]

What would be the best place to start?

If you are working at a University (or if there is one near where you are) why not try to find a lab/group that can use some programming help. You can learn biology from them and in turn offer them programming (bioinformatics) help. As an IT person this would be simple for you (than the other way around).

Working on a real-world problem that someone has right now may be more useful for you than diving into textbooks/web sites.

[krobison]

Some of the less dry & boring reads:

Algorithms on Strings, Trees and Sequences: Computer Science and Computational Biology

Biological Sequence Analysis: Probabilistic Models of Proteins and Nucleic Acids

An Eulerian path approach to DNA fragment assembly.

I agree it is useful to work on real world problems, but it is also crucial to your success that you learn the language of the field and some of the common problems & solutions.