What’s New in Pubmed? Nothing!

There’s a long-standing problem with trying to follow the literature on MSCs. The more proper name is “multipotent stromal cells“, whereas the name many people still use is “mesenchymal stem cells”. There’s also plenty of literature, on what is presumably the same cells, that uses “marrow stromal cells”.

To try to get around this problem in the past, I used the MeSH Browser to find the preferred term for relevant categories, searched on that, and created a feed from those results. That quit working about a month ago. According to the RSS feed for “Multipotent Stem Cells”[mh] OR “Adult Stem Cells”[mh] OR “Mesenchymal Stem Cells”[mh], nothing is new since the first week of November. I see today there’s an announcement from Pubmed about their update to MeSH for 2008, so I’m guessing there may be some connection between the two.

In the new hierarchy, multipotent stem cell, mesenchymal stem cell, and adult stem cell are all under stem cell as distinct categories. Multipotent stem cell is defined as “specialized stem cells that are committed to give rise to cells that have a particular function” whereas mesenchymal stem cell is defined as “Cells that can develop into distinct mesenchymal tissue” and Adult stem cell is defined as “Cells with high proliferative and self renewal capacities derived from adults.” I don’t know how they derived these categories, but clearly they aren’t non-overlapping categories.

Here’s what I get when I run various queries:

MeSH no MeSH
mesenchymal stem cell 2180 3386
multipotent stem cell 708 857
Adult stem cell 261 642
all 3 with AND 3 6
all 3 with OR 3025 4671
difference of OR and summed individual results: 124 214

So both MeSH and keyword queries lose some results when combined, and though MeSH queries lose less, they lose less of a smaller set that doesn’t contain as up-to-date results. I guess I’ll have to subscribe to each feed separately, or maybe mash them together with Yahoo Pipes or something.

3 thoughts on “What’s New in Pubmed? Nothing!

  1. I think you are fundamentally wrong about “losing” results. As an article can be in both categories, the result of ORing all hits can of course be less than the numbers added.
    Using AND does make no sense at all!
    The problem with new and old MESH terms is that old terms are not updated.
    In general: if you want to search the latest content (which has not been indexed with a MESH) or search more sensitive, don’t use a MESH. If you want to search specifically and don’t care about the latest stuff: use MESH.

  2. Thanks, Martin, for that information. The issue I was trying to resolve is whether or not the categories were in fact considered to be non-overlapping. Because an OR search of all three yields less than the sum of OR searches of the three individually, that proves that the categories are overlapping to a small degree. The purpose of this exercise was to figure out how Pubmed uses the different categories and how MeSH searches compare with non-MeSH, since my previous saved searches began to yield 0 results.

    The original goal was to find a Pubmed query that returned all results containing any of the three more or less synonymous terms multipotent stromal cells, mesenchymal stem cells, and marrow stromal cells. The expected behavior was that a search for any of the three terms using MeSH would return an identical set of results. It turns out that only one is a MeSH term, but I also found, as you mention, that using MeSH terms doesn’t return the latest content, so I’ve stopped using them for my “What’s New” searches.

  3. Well, assigning mesh terms is done manually – and this process takes time. I’ve done some work on this and found that indexing the big 5 journals in the field of cardiovascular medicine (i.e. a block buster topic) can easily take 2 weeks.

Leave a Reply

Your email address will not be published. Required fields are marked *