I recently changed the way I approach reading an article, thanks to advice from a science communications class I'm taking. I now start with the conclusions, then intro, figures + captions, and abstract last to make sure I understand the flow of the paper. If it's something where I need to understand the methods in detail, I'll usually then go back and read the whole paper in order.
Starting with the conclusion seemed odd to me at first but has been really helpful and time-saving. I highly recommend at least trying it.
I am still working on my annotations system. Right now I've got a list of things it might be useful to know about a paper; depending on what I'm using any given paper for (background info in lit review section of my article, get ideas about experimental methods, get a sense of sexy new problems in my field) I'll take down notes on some or all of them. These all live in a single big Word file, so I can just search if I need something; as backup and for something more organized, I also paste the notes into the relevant Zotero entry.
I came up with the "useful things" list using stuff other people have posted about their literature review strategies, then just tweaked it for the ways I usually think about papers.
Title, Authors, Year, Journal:
Purpose/Hypothesis/Research Question:
Study organism:
Sample (size, M/F ratio, age, how they were selected):
Design:
Confounding variables:
Analysis:
Results (with specific):
Overall conclusion:
Limitations:
My objections:
Links to my research:
Directions I can take this:
Other notes:
As far as journal subscriptions, I have Google Scholar alerts saved for a few keyword searches and scientists whose work I like to follow. I also subscribe to an RSS feed of PNAS, and skim the biweekly titles for anything related to my interests. This is probably something I will expand a little more in the future, either through more keyword searches or TOC subscriptions.