Sunday, January 13, 2013

Navigating (and drowning in) the flood of PLoS ONE journal articles

I love PLoS ONE--both the mission of the journal and much of the science that is published there--and for the large part I love the new website redesign. But one thing I'm definitely not feeling is the revamped e-mail alert system.

I will admit it up front: I still abide by some old skool methods for discovering relevant literature. Every week, I pour through the Table of Contents and early article alerts from my favorite journals,  neatly delivered to my inbox. Twitter also helps me find a lot of literature, but I find it to be more of a stochastic and unpredictable method (particularly for weeks where my time for social media is limited due to a heavy workload or lots of travel). Plus, being on Pacific Time puts me off kilter with the rest of the world--relevant information is very easily buried in my Tweet stream, even on days when I am looking. So I stick to my e-mail alerts to make sure I don't miss any exciting new science.

Up until a month or so ago, the PLoS ONE e-mail alerts were a behemoth, but they were manageable. The HTML e-mail was nicely formatted with embedded links to a list of articles in fairly specific subject areas, such as "Marine and Aquatic Sciences" and "Evolution and Ecology". It would take a couple of minutes to scan through these sub-categories, but for the most part it was a pretty good way to filter out the research areas which were most certainly not relevant to you. Also, many articles were placed into multiple categories, so an environmental metagenome study using novel analysis methods would be listed under the subject headings for "Computational Biology" and "Evolution and Ecology".

So much to my dismay, I've been going through my holiday-induced backlog of journal alerts and was horrified by the new format for PLoS ONE Table of Contents:

The subject headings that were formerly useful for me have now been completely condensed into the very broad subject headings "Biology and Life Sciences" and "Environmental Sciences and Ecology". Worse, each of these subheading seems to contain a ridiculous number of articles (I didn't count how many, but I was scrolling for a looooong time before I reached the end of the subsection). And it also seems like I need to be looking through both of the above-mentioned subject headings: there were a few relevant articles peppered amongst lots of non-useful literature in each subheading.

I don't have the time (nor do I want to make the time) to scroll through lots of irrelevant scientific literature essentially looking for a needle in a haystack. So I took the advice of the yellow banner and went to create a custom alert on the PLoS ONE website. Frustratingly, there is no way just to look for new articles within in a defined subset of subject areas. You have to include a search term, which immediately narrows your search window. I tried just doing a simple search for "metagenomics", but I was getting a lot of biomedical/clinical articles amongst the interesting ones (and I didn't want to scroll through all 963 articles). Plus, I'm paranoid that my search term isn't catching all the articles that I would want to see. I tried filtering down the articles to a more manageable set, but my attempts did not go over very well:

My final gripe is the subject categories themselves. The checklist of subject terms initially presented under the Advanced Search function is different from the larger list of subject terms listed under "filter by subject area". Are the "filter by subject area" search terms defined based on the articles themselves? I have no idea. I also have no idea what half of the subject terms mean. There's a subject term called "Sequence Analysis" and also one called "Research and Analysis methods" - could/should an article overlap these two terms, or are they referring to two distinctly different things? In my mind these categories seem a bit too vague and redundant to be much use for users. Subject terms also have some glaring errors--there is no "Ecology" category at all!

In the end, I basically gave up. I'm going to begrudgingly go back to that monster of an e-mail Table of Contents. 

I'm a firm believer in intuitive web interfaces with powerful user functionality--I don't think any researchers should have to work this hard to complete what is essentially a very simple (and very common) task. It also in the best interest of journals (and authors) to have their work easily accessible--the articles I'm downloading today may result in future citations, blog posts, social media sharing, etc.

So my pleas for PLoS ONE:

  • Bring back the old e-mail alert format! Or even better, a new revamped format with even more useful subject categories.
  • Consolidate and streamline the subject terms - make them consistent between search interfaces, and specific enough that the meaning of each term is obvious. Ideally, each article would have something like Mendeley tags that would function as searchable keywords. If I liked a particular article, there would be a way to view articles with similar keywords--kind of like "Customers who bought item X also looked at these items" on

I know these type of fixes won't necessarily be easy - I don't know how PLoS ONE organizes its article databases, and the things I'm suggesting might require a significant amount of coding and/or manual curation to implement. But I do think this type of organization is imperative for the long-term business model of the journals. Keep the scientists happy!


  1. Great feedback. If I was at PLoS ONE, I'd be thankful for members like yourself who add so much value with their feedback.

    I have a question which you may have answered in a prior post. What would be the "perfect world" scenario for consumption and publishing of scientific journal work?

    Not sure if you'll be at ScienceOnline, but if so, see you there!

  2. Email?

    I'm an RSS fiend myself. Now the PLOS ONE stream at first might seem to be a deluge but if you apply some conservative keyword filters to the 'full-fat' RSS feed, you can tame the firehose of science down to a manageable and relevant trickle in my experience (although this might depend on how broad and [keyword]definable your research interests are, for me it's not too hard).

    I've got a post on one way of doing it here: describing the use of Yahoo Pipes for such a purpose (but I'm sure there's also other ways and means).

  3. Cyrus - I don't know about a "perfect world" - I think these type of things differ very much from person to person. Ideally I think there should be a variety of solutions that work well, and each person can customize/combine their use of different tools to suit their own personal preferences. Kind of like social media tools - some people prefer long-form blogging, while others prefer shot-form Twitter interactions. I'll be at #scio13 - looking forward to catching up!

    Ross - Thanks so much for posting this tip here. I love RSS feeds too but I often find them overwhelming and I tend to ignore them more than my e-mail. I've tried so many times to keep up with them but I always fail. Usually I use RSS for blog feeds rather than journal alerts, but perhaps you've convinced me that I should give Yahoo pipes a shot.