I have a strong mathematical and computer background (also with machine learning) and would like to switch to bioinformatics. I know a fair amount of biology. But I find it hard to find real world practical training in this field. So I would like to try to replicate some bioinformatics papers (with their provided data). Good beginner papers for me would be shorter ones where the biological problem is well stated and the problem is also fitted in to the bigger scope (or is just easily understandable). Do you have any papers to recommend?
1
-
$\begingroup$ I'm sorry, but this really isn't a good place to ask this sort of broad, open ended question. We focus on specific problems here with specific answers. Asking for a list of papers isn't really on topic. You might have better luck on biostars.org $\endgroup$– terdon ♦Commented Dec 25, 2019 at 15:30
Add a comment
|
1 Answer
$\begingroup$
$\endgroup$
2
Bioinformatics is a big and very heterogeneous field with a lot of variation, so it's hard to recommend something that covers everything. That said, I think the best you'll do is to build familiarity with some of the common tools with some canned analyses.
One possible option would be some of the scRNA-seq workflows from the Pachter group, they seem to be fairly conscientious about documentation. Here is one such canned analysis, here is the paper, here is a blog post about it that suggests next steps and discusses at a high level.
scRNA-seq is just what's been trendy for a few years, next year it will be something different. But it at least gets you familiar with some of the file formats and concepts such as alignment, etc.
Hope that helps.
-
$\begingroup$ Thanks for your answer! I'm not looking for something that covers bioinformatics, I'm looking for one or two bioinformatics papers which are replicate-able for someone that has started scratching the field. Most bioinformatics books mostly covers the different file-types and transformation between them, I want something with a little more substance but not to many layers (deep) in the biology part. e.g using machine learning to figure out the genes that are responsible for a particular genetic disease. $\endgroup$– NatanaelCommented Dec 27, 2019 at 12:04
-
$\begingroup$ Well, then the linked analysis should help. It gets you from the raw data to data matrices, and points to clustering methods on those data, without getting too worried about the biology. Like many fields, that's what bioinformatics is, is composing sets of tools to get to datasets that you can plug into ML or whatever. There are many different ways of doing it, and many problems that can be studied by doing so. The linked analysis is one set of methods applied to one problem. Most of the "substance" happens upstream of the clean dataset, as with all analyses (even ignoring mechanism). $\endgroup$ Commented Dec 28, 2019 at 3:50