3 Comments

Last tonight I was browsing Occam’s Typewriter, one of my favourite science blogging collectives.[1]

Frank’s article, The challenge of going beyond, leads off from two people independently remarking to him that they thought that electronic journals have resulted the loss of serendipity in their reading. The article and comments are well worth reading; I encourage my readers to head on over if you haven’t already.

Contrary to the people the commented to Frank, I don’t think that electronic journals have changed how I capture serendipitous aspects of my reading all that much. If anything they’ve made it a little easier, if a lot more distracting!

The main difference I can think of is that I now do what I once did through print editions using the on-line editions. Otherwise I do basically the same thing. (Boring? Sad old habits?!)

I’d like to think that I’m not different from most scientists in having a small(ish!) collection of favourite journals that I scan every week, or month as the case may be. Journals that are loosely in my areas of interest, but also contain enough tangential papers that I pick up the odd new development outside of what I happened to be after at the time.

For what little it’s worth, my list includes the ‘obvious’ journals for anyone interested in molecular biology (Nature, Science, P.N.A.S, Cell, etc.) and some from each of genetics, biophysics and bioinformatics, with a few on proteins (my old haunting ground). Rounded out with a dash of neuroscience (a newer interest), you pretty much have my lot. It’s plenty to cover and wide enough that I can’t read everything that strikes me as interesting. Then again I find even reading the abstracts or summaries clues me in a little to what is happening in other areas.

I’m a computational biologist. I sometimes wonder if experimental biologists think I mainly read computational methods and whatnot. In practice, I read more experimental biology papers than computational ones. Obviously I read the main important computational developments and key analysis papers–I couldn’t ignore them–but I have a long-standing wariness of algorithms (largely) developed first, then ‘fitted’ to data after-the-fact.[2] I prefer to try understand what biological questions might usefully be asked using computational methods, then with that in hand look to see what existing methods might be suitable in addressing them, or develop my own algorithms to tackle them (more fun…!). Basically, you have to have a good understanding of the biology before developing an algorithm or analysis strategy.

I leave keyword search-based approach for when I’m trying to gather in all the papers on a particular topic, when writing a review paper or grant application. (An exception is a few keyword based feeds I get on long-standing topics of interest that can be captured well in a few words.) They’re useful, but I don’t feel they do much for serendipity. For me, anyway.

On rare occasions I will pick one of the larger journals series or publishers and simply browse the journals they offer. I have to admit I rarely have the time for that luxury these days, but it still does happen from time to time! I discovered one of my favourite ‘little’ journals this way – Epigenetics and Chromatin.

One difference that Heather pointed out, that I agree with now that she mentions it, is the ‘related articles’ suggestions some journals offer. On the same note, the ‘cited by’ references can be very useful. Printed cross-indexing series once served a similar purpose, but where much, much slower and lagged behind the literature.

I used to read NewScientist (American Scientist sounds good, too), but I have to admit I’ve lost interest in them. Aside from that a few articles seem to be a bit wayward, these days I read blogs for much the same sort of content.

This paragraph from Frank’s round-up comment strikes a chord:

My boss commented to me that ’it’s a state of mind as much as anything. If you are the kind of person who likes to read around a subject and even outside it, you will do so.’ I think there is a lot of truth in that. I remember years ago seeing an article recommending a good reading habit for researchers as being to read one textbook in your broad discipline every year, regularly read review articles on the fringes of your field, and of course read the primary literature of your field.

Reading the literature can be hard work–what isn’t–but it’s an on-going learning experience. Isn’t that what we’re all after?[3]

Footnotes

[1] If any of the Occam’s Typewriter crowd find themselves over this way: you know someone likes your stuff! (Maybe I’m biased from having spent my Ph.D. years in England?)

[2] Generally speaking, either this isn’t the problem it was, or I’m better at avoiding them as I grow older!

[3] Of course I’d want to do some actual science.


Other musings on Code for life:

Where do good ideas come from?

Sinclair ZX envy

On alternatives to academic careers and “letting go”

The roots of bioinformatics

Rain, sleet, snow, music and science blogs