The main way scientists share their findings is through publishing scientific research articles. Reading this literature has tremendous benefits – everything from staying up-to-date on the field, to helping you design experiments, to helping you avoid trying certain experiments, to allowing you to connect ideas and come up with wild, but maybe legit, theories. But reading this literature can also be really hard. So hopefully this post can help you read with purpose and get the most out of the articles you read.
note: the text is an adaptation of a mashup of stuff from past posts (so sorry for formatting weirdness) and there are link to those posts at the bottom
This is probably one of the most important things to stress – each person will have their own style that works best for them and the best way to get more comfortable reading papers is to read a lot of them.
When it comes to approaching papers you have to think a bit about what your aims are. Broad level understanding vs nitty-gritty etc. Sometimes you can get super sucked into some particular aspect of a paper and trying to interpret a little thing that you waste a bunch of time and/or lose site of the bigger picture. It’s good to try to get a broad overview of the paper – I like to skim through and look briefly at the figures to see what type of experiments they’ve done and what to expect. And I read the subsection headers before diving in
If it’s something important, then I will go in depth and even pull up the supplementals, etc.. It’s basically impossible to do that for every paper you come across, but if you’re basing your project off of something or citing something, you want to make sure you understand it and the papers conclusions are well-supported by their evidence
I definitely recommend trying to really understand what each figure is saying and why it’s included. Even if you aren’t looking for the nitty-gritty details, negative controls, uncrossed gels, etc. you can find in the supplemental materials, there’s still a reason to check it out. Look in the supplemental to see if there’s a model figure. for some reason, papers tend to relegate this most helpful figure to the supplemental, usually at the very very end, but having that graphic layout of their conclusions will help you see the bigger picture and try to figure out how each piece of evidence fits (or conflicts with) their story
A lot of people skip the intro but I find the jntro really helpful – and the methods (but part of that’s just the geek in me). Methods can definitely be helpful if you’re looking about ideas for your own experiments, but Methods can also be frustrating because of that citing papers that cite papers that cite papers… that I mentioned before.
The intro often references review articles – if I’m new to the field or topic I tend to try to check out the most recent review article they cite (not that the most recent is always the best but I want to make sure that I’m up-to-date. they’ll often cite the same older one and then that might get me to look up the earlier one too)
One thing I found really valuable was taking a more “historical” dive into my field, going back to the early 2000’s when the protein I work on was first discovered, and seeing how the knowledge built up. It’s easy to get swarmed by new papers, but this foundational stuff is so valuable and helps to see various strategies that’ve been used. I’m always afraid when reading older papers that I’ll be using “outdated” information, but I’m using a browser extension called scite_ that helps me see where it’s been cited & context so I know if it’s been validated, refuted, (not a paid endorsement, I’ve just found it helpful).
I keep notes on what I read – usually I copy & paste the citation & link into my e-lab notebook and then make bullet points (doing it in the e-format makes it easy to search). I make sure to copy & paste all the info on authors, journal, year, etc. because often people will refer to a paper by one of the authors or the year and journal (they often don’t remember the paper name and you can’t blame them!). The most important authors to note are the first author & the last author. In my field at least, the first author (sometimes there are co-first-authors) is usually the person who did the wet work (e.g. the grad student) and the last author is the PI. But it varies from field to field and sometimes is just alphabetical, etc.
my advice on what else to write down:
- start by writing 1-2 sentences summarizing the key findings & your impression
- then take bullet point notes on
- what types of experiments did they use?
- how strong was the evidence & what were the strongest/weakest points?
- when applicable – how does it potentially fit into YOUR research?
- take more detailed notes on experiments that might directly impact your work &/or change your thinking
I also note when people are the first to name or discover or propose something so I remember that this is the “original source” if I want to cite that thing in my paper. If I’m not sure about first-ness, I’ve also been using the scite_ tool to see what others have cited the paper for. Speaking of citing, I use a free reference manager and citation generator called Mendeley. It has a browser extension that lets you add articles to your library from the article page. It also has an interconnected desktop version and a Word plugin so you can insert citations from your Mendeley library into your Word document. https://www.mendeley.com/reference-management/reference-manager
more on reference management here: https://bit.ly/referencesreference
Sometimes I come across papers that I don’t have access to because they’re “paywalled” (they want your money). You can find out more about this here: http://bit.ly/openaccesspreprints
At schools and research institutions (or your local public library library) this happens less often – the library subscribes to various journals, so *you* have access to their articles (you might need to be on the campus’ WiFi or VPN) but that doesn’t mean that everyone does. If you want to know if everyone does (i.e. it’s open access) look for an orange unlocked lock icon. You can also install a browser extension called unpaywall which will put a green lock icon on articles that have a (legal) freely available version somewhere on the web and direct you to it. https://unpaywall.org/products/extension
If you come up against a paywalled article you really want, ask a librarian! They can often get the article through “InterLibrary Loan (ILL)” (they can do that for books too). You can also try e-mailing the paper’s contact author (papers usually provide an email address). Or asking people on Twitter (which I’ve successfully done on multiple occasions).
As for keeping up with the newer literature, I use a website called Meta to keep up to date so I don’t miss things & I check that daily. Time permitting, I usually read after I get done with my experiments (but by that time I’m often already really exhausted so not ideal). I try to deep-read papers that seem especially relevant to my project and usually skim-read-skim-read others. Lately I’ve been trying to deep-read more because I’m working on my own paper and trying to get a better sense of how I should be formatting things & what’s good/bad. https://scite.ai/
Speaking of bad, a browser extension from PubPeer can help you weed out really bad papers. PubPeer is a site where people can comment on papers. The comments can be good or bad, but most of the time these comments are drawing concerns about figures potentially containing altered images, etc. You can install a browser extension that will give a popup if articles on a page you’re on have comments on PubPeer. It will also tell you if articles have been amended or retracted. https://pubpeer.com/static/extensions
I’m not sure if this “advice” is helpful or not but it’s how I approach things. It definitely takes a lot of practice and is acquired skill. And even experts have a hard time reading papers, especially those not in their sub field So don’t feel bad or discouraged if you’re struggling. Some of it is the authors’ fault anyway for not making their points more clear. People seem to think that serious science = dry, boring, dense, and jargonny so scientific papers can be really inaccessible (even when open access!)
more on the layout of science articles: http://bit.ly/readingsciencearticles
if you want to see an example of walking through a journal article: http://bit.ly/averyarticle