Hacker News new | past | comments | ask | show | jobs | submit login
Rosalind is a platform for learning bioinformatics through problem solving. (rosalind.info)
109 points by nkurz on Nov 9, 2012 | hide | past | web | favorite | 22 comments

I wonder what are the potential career choices in bioinformatics for a computer science major who hasn't taken many biology/chemistry/genetics courses. Any insight?

I work on a bioinformatics team in a research environment and have no bio coursework/background. There is a ton of standard IT work here.

I do mainly data integration and standard reporting stuff. For example, reports for "where is my sample in the pipeline, has the fastq been generated, where are the variant call results?" But being interested in the bio/genetics stuff is a huge plus. Hackers in the space seem welcome and needed. You gotta fight the credential hierarchy some - you clearly don't matter as much as the MD/PhD, but sometimes they should listen to you, like when, say, it comes to the programming language choice you make for the app they will use. That said, I sneak in what I think is best for the task and if that means Ruby or Clojure instead of Java, results are what matters in the end.

I work at Counsyl, a pre-pregnancy genetic testing company, and I can confirm that there is a ton of work for people not versed in bio. Everything from assay & sequencing visualization tools to more traditional IT challenges like "How do we manage the huge amount of data coming off the seq machines?" -- which is terabytes per day.

I don't have a bio background so I work on software tools and automation to support the rest of the company. Counsyl has huge respect for all levels of science and engineering, so you don't face a credential hierarchy everywhere (but Counsyl is probably unique in this area). We have a strong solve-problems-with-software culture that any dev would be comfortable with.

If the gp is interested, check out https://www.counsyl.com/jobs/

A key activity in bioinformatics is developing software tools for analyzing and interpreting data. If you have no knowledge of biology, you could still work as as soft engg. in this field.

If you are willing to learn basic molecular biology (structure of DNA, central dogma, etc.), you can get a job as a bioinformatician, bioinformatics analyst, bioinformatics engg, etc. where you could be analyzing biomedical datasets.

Isn't bioinformatics like biology in that the minimum qualification to be more than someone else's staff/glorified lab tech is a Ph.D.?

No. I sit side-by-side with PhDs in bioinformatics and am useful because I can write crazy SQL, can munge data at the cmd line and with ruby/clojure, and simply don't care what format data comes to me in. I have some preferences, but I'll work with whatever - many of the bioinformatics PhDs will know a little scripting (Perl or R) but are more valuable doing other work rather than pounding out parsing scripts and doing data transformations.

We do work that winds up in people's actual medical records - meaning that genomics analysis we do influences/directs actual medical treatment for patients.

That said, I wish I knew more biology and statistics, which is just interesting to me. I'd feel a little "key" to the work rather than a "nice" addition to the team.

Forgive me for being a status/autonomy whore but it sounds like you are a respected, professional, “part of the team”. Neither money nor more importa tly prestige flows that way.

I have absolutely no doubt that the work you are doing is useful, relevant and a genuine contribution to society and the enterprise of science. But how many people who are “just” programmers or “just” bioinformaticians, no Ph.D. end up as first, second or third authors on papers?

That is the coin of the realm, right? It certainly is in Chemistry, Biology and Big Lab/Group Physics, yeah?

No worries, I think you have the right read on the prestige issue. Of course in an academic research center, no one makes any money and prestige is reserved for grant-winners and first authors in Nature.

Publications would be hard as a non-PhD. If you wanted to focus on that, I have found that PIs welcome outsiders into their research groups. I'm certain you could put yourself in a position to publish with the same amount of credit as a grad student on a team (not much?). Especially with good data hacking skills, someone will be willing to help you scratch that publication itch. I think it would be like having a major side-project in addition to a fulltime job though.

Oddly enough, even though the work I do here is less technical than what I have done in the private sector, I seem to perceive a slight up-tick in respect from enterprise software groups.

Papers can be published about tools that are developed. For example, NAR has a web server issue every year. http://nar.oxfordjournals.org/content/40/W1.toc

I think they 'd be great if you are willing to acquire the biological background. Due to its descriptive nature, that could be easy-peasy.

I'd like to see a similar site for problem solving in computational chemistry and physics.

Anyone know of any other good online ressources to learn more about bioinformatics? I think its a fascinating topic and have been meaning to learn more about it for a while now.

Don't start with organic chemistry or genetics unless you have a 5-year plan. Basic understanding of molecular biology (structure of DNA, central dogma, etc.) will be enough to get started. Finally, here is a book (1) that describes many algorithmic problems in bioinformatics that a techie could relate to.

[1] http://www.amazon.com/Introduction-Bioinformatics-Algorithms...

Coursera seems to have s course on Genome Science: https://www.coursera.org/course/genomescience

You have to start with organic chemistry and genetics. If you don't, you won't have a good understanding of what's going on.

Bioinformatics itself is huge, there are many many areas - protein folding, protein interaction, DNA/RNA interaction, all kinds of taxonomies, evolution, gene mapping, systems biology, etc etc etc.

I strongly suggest you pick something particular, because, otherwise, it's easy to waste years on learning many topics and not have any deep progress on anything in particular.

Nice. But I can't wget the datasets. There's magic behind the "Download Dataset" button for each of the problem sets - that's quite annoying programming via an ssh connection.

This is because the datasets are unique for each user, so you need to login first, and then download the dataset.

Sure, but that's no reason to not make the URLs available for logged in users.

A lot of the early ones are good for testing your linux-fu I'm finding :) Very cool & interesting

Similar to Project Euler except for bioinformatics

Good Lord, these are like fun little programming potato chips.

Update: Aaaaaand then they get to be fun big potato chips.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact