Eldan Goldenberg's lab notebook

Keeping references organised

Because I'm starting on a somewhat new research track, I'm finding myself spending the majority of my work time reading. This brings me back to a question that I think about every time I'm in this sort of position, and never seem to have a satisfactory answer to: what is the best way to keep references organised?

The way I see it, there are two challenges involved in this: keeping track of what one has read, and keeping track of where one read it. The first part is a general problem with learning new things, and while I don't think there's any magical solution, I also don't think it's all that hard provided one pays attention and thinks through how each paper relates to what one already knows. It's really the second I'm interested in, because obviously it's impossible to cite work appropriately if, as I often find, one only remembers I read a paper about this once without being able to pin down which paper it was.

Behind the cut I'll write a little about how I try to manage this, but I don't believe my own approach is optimal, so I'm very interested to hear from my 3 readers about how you do it.

My process involves two strands: using software to keep all the bibliographic information recorded indefinitely, and writing documents to keep track of what ideas were in which papers.

The only software I use for this is an online service called CiteULike. It allows me to maintain a library consisting of reference information, [usually] abstracts and [occasionally] my own notes about what I found salient in a given article. I can sort/filter the list by priority in the 'to-read' queue or by authors or tags, which is great for keeping track of papers I have as yet to read, or for tracking down a specific paper when I remember who it was by, but doesn't help with the problem of remembering in which paper I read about a particular idea or set of experiments.

My only solution to this problem is a low-tech one: I'm incrementally writing a literature survey to which I add each paper, with a sentence or two to give context, as I read it. This isn't the sort of lit review I would ever give someone else to read, because it's very redundantly structured (e.g. the same papers are listed in one section filed by model organism, in another filed by computational model type, and in a third by what I think I learned from them) and as patchily written as you would expect from something to which a few sentences are added each day. However, if I were to one day write a computer models of gene regulatory networks lit review for other peoples' consumption the task would be very much accelerated by having this resource, and more importantly I'm hoping that when it comes to writing introduction and discussion sections of papers and eventually a thesis this will save me a great deal of time.

This process basically works for me, but I'm really not convinced it's the best way possible, so readers: what works for you?