When I was but a lad, my siblings and I used to accuse my father of knowing everything, a charge he would always deny, observing that even excluding those things which are not known to anyone or are secret, and restricting ourselves to academic knowledge, there is more to know than any one person or any thousand people could possibly know.
To see why this is so, consider the sheer volume of scientific literature being produced. ISI Web of Knowledge, an online tool used primarily by scientists for finding scientific literature relevant to their work, indexes the contents of over 23,000 academic and scientific journals. Many lesser known, newer or otherwise less main-stream or traditional journals are not indexed at all. One needs to go to other databases to find information published in books, or in dissertations, or, or, or.
Like my father (as I am in most things), I find myself far too mortal to know any meaningful fraction of anything, even if we restrict ourselves just to academic biology. I'd say there are roughly 8000 peer-reviewed journals in which biological work is regularly published, and if you add up all the papers I skim through, it is probably the equivalent number of pages of the output of two or three of these. The papers I read in detail if assembled together would surely make up much less than the output of a single journal. And I put more time into reading and therefore less into writing than is optimal for my career.
One result of this is that I frequently find out that there exist thriving sub-disciplines of biology of which I have almost no knowledge. For example, only last week I for the first time heard the word "metabolomics." Google Scholar lists over 6000 papers in the last year on this rapidly expanding field about which I know no more than I could guess based on the name.
My knowledge of transcriptomics was quite as absent three years ago. Transcriptomics is the study of RNAs in the cell, generally in the context of gene expression patterns. I became interested in transcriptomics because I proposed that mortality risk during embryonic development would be highest at those stages at which gene expression patterns were changing fastest. I was a doctoral student at the time, and my adviser asked me if there was any way of testing this idea. I had to admit I didn't know if it was possible to test because I didn't know enough about the field I have since learned is called transcriptomics. The answer is that yes, there is a way of testing the idea, but it will cost a couple of hundred thousand dollars, and require collaboration with people who read different journals than I do. I will never be an expert in transcriptomics, but I can find a colleague who is, but has little knowledge of evolutionary demography, and invite him to collaborate on a project that combines our expertise. And this is why science can be a somewhat unified pursuit despite having far more product than one person can read even the titles of.
I referred in my last post to one of my advisers at Berkeley, and one of my all around favorite human beings, Ron Lee. Ron would always advise me to think about my relative advantage, by which he meant I shouldn't just work on the most interesting or best questions, as there are far too many. Rather, I should choose among them by considering which questions I was better placed, given my strengths and resource, to answer than was anyone else likely to work on the question. Ron's advice has always served me well, and so I do consider this before starting any project. Frequently, as with this developmental transcriptomics and demography project, my relative advantage arises from the fact that I am combining two fields separated enough that no one else is likely to ask the question any time soon. Evolutionary demography and developmental biology do not, as a rule, talk to each other. Many fields of biology have almost no communication with each other, leaving vast unexplored interdisciplinary territories (unless that is all in some set of journals I haven't come across yet).