Comments on: Why Bioinformaticians have to grin and bear it!

By: Wilf

Wilf — Tue, 09 Dec 2008 16:20:25 +0000

The biologists are always ahead of informatics because they’re in the business of creating innovative procedures that won’t necessarily conform to existing data structures. They do know spreadsheets, which we can deal with fairly easily and which they can use as an external data format until we can integrate it (or its abstraction) with our system. The essential elements are good communications and fostering team unity.

By: harijay

harijay — Mon, 17 Nov 2008 19:52:19 +0000

Oh and one more thing..I think the “Saunders principle” is a great meme ..I plan to use it the next time I have to talk about the importance of structure and data

By: harijay

harijay — Mon, 17 Nov 2008 19:49:48 +0000

Thanks Neil for your excellent comment. I couldn’t have said it better . In the end we all benefit by consistency and standardization.
What we do is difficult enough and every skill we apply to make our life easier cannot be underestimated, whether it comes from better designed experiments and data formats or more lucid specs and bioinformatics workflows.

I really respect bench-scientists turned bioinformaticians and think you fortunately bear the burden of bridging the divide . Heres to more conversations that help educate both sides ..and I am with you a 100% about word attachments ..every carrot does need a stick at the other end…thats the only way we will all learn and benefit

By: Neil

Neil — Mon, 17 Nov 2008 13:58:32 +0000

I’ve been meaning to comment on this for ages, but wanted to get through the podcast first. I think you guys covered it though; in particular I agree with Deepak that “everyone is right, everyone is at fault”.

I don’t expect experimentalists to become computer programmers, or to care very much about data formats. What I did expect, for a while, is that they’d be interested in anything that made their lives easier or their research more efficient. Experience has shown that if this involves learning even a modicum of new computer skills, it’s usually a vain expectation, with rare exceptions. When I “complain” about biologists, it’s only out of sadness that they are missing out on all the wonderful tools that make my own work easier and more enjoyable. And I speak as a wet lab guy turned bioinformatician, so I’ve seen both sides of the coin.

I don’t know how to overcome this reluctance to consider anything that lies outside of a narrow band of education and skills. It’s extremely prevalent in academia, at least in Australia and leaves me quite depressed at times. But as you say, we’re all wired differently – I just have to accept that people make their own choices as to what’s a good use of their time.

That said, the next person who send me sequences as a Word file is going to find it back in their inbox with a short note as to why this is inappropriate

By: Deepak

Deepak — Thu, 13 Nov 2008 02:21:00 +0000

Jan and I agree on the consistency part. It’s the most important part. no consistency, no workflows

By: Coast to Coast Bio Podcast » Blog Archive » Coast to Coast Bio podcast #3

Wed, 12 Nov 2008 19:26:34 +0000

[...] and the Saunders principle wherein Bioinformaticians struggle with inconsistent formats , prompting code-itch (Hari) to complain that Bioinformaticians have to grin and bear it! [...]

By: harijay

harijay — Mon, 10 Nov 2008 15:18:23 +0000

I wholeheartedly agree Jandot , its very annoying to see inconsistencies and many a time I too find myself “improving” my data reporting by ad-hoc changes to the format or level of reporting. Somehow I think these problems are more plentiful in academic research . But I am convinced , that the way out is to gently coax everyday experimentalists the importance of structured data. At the risk of stating the obvious, if experimentalists were aware of the purpose of xml or json or rdf etc collaborations would definitely benefit

By: jandot

jandot — Mon, 10 Nov 2008 14:59:16 +0000

I _completely_ agree with your remark that “experiment design does not extend to the data storage or management level”.

Apart from that (and being a bioinformatician who has do all the data massaging afterwards): we cannot always ask wet-lab experimenters to provide their results in a fixed format because the type of results is not known yet. Often it is (e.g. running standard PCRs or resequencing), but not always.

What I mainly want from my data providers (read: the lab people generating the data) is that they are *consistent* in reporting. As long as the files I get are nicely structured, I can get thing done. But if they give me an Excel file with for example one sheet of results per chromosome and each sheet has a different format (even though it represents the exact same type of results) I know I’ll have a bad day/week.

By: Deepak

Deepak — Fri, 07 Nov 2008 15:50:15 +0000

I think we have something to talk about this weekend