R Computing Language

[deleted626763]

Advertisement - Members don't see this ad

Hi -

I'm a double major between Psych and Biology. I took proper mathematical Statistics for the Bio major. The Psych major wanted something directed towards Behavioral Science research specifically, so I took their Statistics for Behavioral Science Research too.

I have historically always used Stata and SSPS in my research positions, but it seems like R is getting to be the wave of the future.

The Behavioral Science Statistics class had notes on the syllabus where you could self-teach R. I did that and feel reasonably competent.

There is an upper-level Bio class that is available for Bio majors but I don't need that is a Biostatistics class that formally teaches R. It's in the name of the class.

Do you think I can just say I am R-competent and have that work for research jobs after graduation? Do you think it would help a whole lot to have the formal class? (I am deciding whether or not to take the formal class.) It's there and a real class will essentially only be an option until I graduate (pretty soon!) and so I want to do it while I can if I should.

I want something paid for my application cycle. I am a bit of a transient research assistant and don't reliably get paid. That's fine right now but it's important to me to find something more stable during my gap year(s). The paid positions are really competitive with only a B.S. in Biology & Psychology and "just ok" experience.

 

[Kardio]

As an autodidact your knowledge of R may have some holes in it. Self teaching is great but we naturally tend to focus on the elements of a subject we are interested in or perceive as important, rather than approaching the material holistically.

You may get a more complete understanding by taking the class (and probably earn an easy ‘A’ for AMCAS while you’re at it).

 

[deleted626763]

As an autodidact your knowledge of R may have some holes in it. Self teaching is great but we naturally tend to focus on the elements of a subject we are interested in or perceive as important, rather than approaching the material holistically.

You may get a more complete understanding by taking the class (and probably earn an easy ‘A’ for AMCAS while you’re at it).

That's an excellent point. Thank you so much for your feedback. 🙂

I will sign up for the class.

 

[voxveritatisetlucis]

Pandas (with matplotlib and sklearn or statsmodels) >> R imho.

But in all seriousness, it can’t hurt to list programming languages that you’re familiar with. In my experience, most companies will give some sort of assessment (ex. Leetcode) to see if your skill level matches what they’re looking for. A formal class might be a good way to gain exposure to less familiar topics

Datacamp is also somewhat decent and kind of fun for brushing up on R topics that you may not be familiar with.

 

[KnightDoc]

Do you think I can just say I am R-competent and have that work for research jobs after graduation? Do you think it would help a whole lot to have the formal class? (I am deciding whether or not to take the formal class.) It's there and a real class will essentially only be an option until I graduate (pretty soon!) and so I want to do it while I can if I should.

As the cycle progresses, I am proud to say that I, too, am becoming very "R-Competent"!

 

[VaultArmitage]

I am a computational biologist by training and love R! However, I think it has several strengths but a lot of weakness. First weakness is speed. R is very slow unless you REALLY know what you're doing. I'm currently profiling some code that I wrote and realizing that a lot of the ways I learned how to do things are incredibly slow. Second weakness is computationally heavy tasks like machine learning, neural nets, etc. Python just has much faster and intuitive packages.

In my (humble) opinion R is really great at three things. Data visualization, statistical analyses, and cleaning data . Need to make a really complicated graph for a paper? Ggplot all the way. Want to figure out an obscure statistical test between two "dirty" datasets? Base R and tidyverse is more than up to the challenge. Want to do some really cool machine learning modeling to uncover previously unheard of gene interactions? I'm going Python 100% of the time.

In all R is a great language when it comes to those three tasks. Anything more complicated I personally think that R code gets messy, slow, and unintuitive. If you feel competent in R then I would focus on Python/Rust/Java/etc. C++ is more for software development and things that require insane speed and memory usage. But I will fully admit that I am biased against C++ because I have a pretty good grasp of it but not enough to make the annoying syntax and memory leaks worth it.

On an unrelated I think there are many broadly related "computer science" topics that are immensely helpful to researchers. One is cybersecurity and infosec. I am amazed when I meet very intelligent computational researchers that didn't know they need to change the default root password on a linux server with 256GB of RAM. Checkout HackTheBox or Offensive Security for some cool resources.

Anyway sorry for the long response! Hope it wasn't too rambling.

 

[voxveritatisetlucis]

I am a computational biologist by training and love R! However, I think it has several strengths but a lot of weakness. First weakness is speed. R is very slow unless you REALLY know what you're doing. I'm currently profiling some code that I wrote and realizing that a lot of the ways I learned how to do things are incredibly slow. Second weakness is computationally heavy tasks like machine learning, neural nets, etc. Python just has much faster and intuitive packages.

In my (humble) opinion R is really great at three things. Data visualization, statistical analyses, and cleaning data . Need to make a really complicated graph for a paper? Ggplot all the way. Want to figure out an obscure statistical test between two "dirty" datasets? Base R and tidyverse is more than up to the challenge. Want to do some really cool machine learning modeling to uncover previously unheard of gene interactions? I'm going Python 100% of the time.

In all R is a great language when it comes to those three tasks. Anything more complicated I personally think that R code gets messy, slow, and unintuitive. If you feel competent in R then I would focus on Python/Rust/Java/etc. C++ is more for software development and things that require insane speed and memory usage. But I will fully admit that I am biased against C++ because I have a pretty good grasp of it but not enough to make the annoying syntax and memory leaks worth it.

On an unrelated I think there are many broadly related "computer science" topics that are immensely helpful to researchers. One is cybersecurity and infosec. I am amazed when I meet very intelligent computational researchers that didn't know they need to change the default root password on a linux server with 256GB of RAM. Checkout HackTheBox or Offensive Security for some cool resources.

Anyway sorry for the long response! Hope it wasn't too rambling.

Which Language would you say is the most commonly used in computational biology and is it the same in most areas of medical research?

 

[VaultArmitage]

Which Language would you say is the most commonly used in computational biology and is it the same in most areas of medical research?

Honestly I would say Python. It's easy to pick up, powerful enough to handle most tasks, and popular enough that people are constantly adding new packages daily. Here's an interesting link:

Top programming languages: Most popular and fastest growing choices for developers

SlashData's latest Developer Nation report looks at the popularity of programming languages and trends from the global developer community.

www.zdnet.com

While it's only looking at trends for developers and we can't necessarily extrapolate I think it highlights a growing trend towards Python as the language of choice for non web development tasks (For that you can't beat JavaScript in my opinion).

I've worked in 3 computational labs across medical specialties and have seen 75% of the code in Python, 25% in R, and the stuff to run that code all scripted in bash, but again that's from my limited experience.

 

[LizzyM]

Long story short, the more of these skills you can build during undergrad and onward, the more useful you'll be to your team wherever that may be.