Thursday, April 29, 2010

Using Mendeley for Long-Distance Mentoring

Last week I received an email from a student who's applying for one of Emory's undergraduate research programs. The SIRE Research Partner Program matches undergrads with faculty working in their areas of interest, teaches them research and communication skills through a series of educational workshops, and provides funding or class credit for their hours as a research assistant. This prospective SIRE trainee wants to work with my adviser. Once she joins the lab, I'll be training her in various laboratory techniques and providing some supervision for her day-to-day activities. I want to do more teaching and research mentoring, so I'm really looking forward to working with her.

In the meantime, she's asked me for a list of papers that she should read to help familiarize herself with the work we do in our lab. The simplest thing would be to email her some citations or PDFs, but I wanted a more interactive way to bring her up to speed. Papers are always easier to manage once you've talked them over at journal club, and it seems unrealistic to expect a rising college sophomore to understand all of the details in these papers without any help. Ideally, we'd get together regularly to discuss the papers and how they relate to her project, but that isn't possible while she's spending the summer with her family in another state.

So, I decided to create a shared collection with her on Mendeley. Using the free Mendeley Desktop software, I can add papers from my own reference library to a collection that I've named "Summer Reading." (This collection is not public; only the two of us have access to it.) My student can read the papers from home on her own copy of Mendeley, allowing her to familiarize herself with a reference manager (an important research skill!) and communicate with me about her reading list.

I attached PDFs to each reference in our shared collection. Then, I used Mendeley's annotation capabilities to highlight areas of particular interest for her project, and left a few comments in the "Tags & Notes" field to help her focus on the most important points in the paper. These annotations are shared along with the PDF, so my student can follow along easily while she's reading the paper at her computer. I've also encouraged her to add her own annotations as she goes along.

After she's had a chance to read, highlight, and add notes to a paper, she can sync Mendeley Desktop and upload the annotations that she made. They'll show up on my computer the next time I sync my reference library. I encouraged her to include any questions that she has about the papers in these annotations. I think the "sticky notes" will be especially useful for this -- instead of having to quote a paper at length, or refer to "page 7, paragraph 2" in a message, she can stick a digital note right next to the relevant portion of the text. I can continue editing the PDF to respond to her questions, and we can go back and forth as needed.

I did have a little trouble while I was setting this up. This is my first time using shared collections on Mendeley, and some of the documentation on their FAQ is a bit sparse, so I had to go by trial and error. The most important point: In order to collaboratively annotate PDFs, all members of a shared collection must open the collection in Mendeley Desktop, click the "Edit Settings" button, and check the boxes for "Upload attached files to shared collection" and "Download attached files from shared collection." Otherwise, PDFs attached to the references won't show up, and annotations made to those PDFs won't be shared with other members of the collection. These are turned off by default, and until I stumbled onto the relevant menu options, I had no idea how to make attached files show up in the collection. I made one of my labmates join a shared collection with me, then ran back and forth between our two computers a few times until I got it working. It seems obvious now, but I thought I'd mention the solution in case other people who want to try this are as clueless as I am.

We'll see how this works over the summer. If written communication proves insufficient for our needs, I may suggest a few Skype chats to talk things over in real time. Anyone else have suggestions for long-distance collaboration / educational tools?

Sunday, April 18, 2010

Troubleshooting Gateway Cloning

After struggling with a particularly annoying problem in the lab, I finally got my experiment to work this weekend! When I crowed about this to my partner, he told me that whenever he figures out the solution to a tricky problem in his line of work, he blogs about it, for the sake of others who might have the same issue. I did my share of Googling while troubleshooting my experiments and didn't find a solution, so I thought this was worth blogging about. The following will make very little sense if you don't do molecular cloning.

I spent almost three months trying to clone my DNA sequence of interest into a mammalian expression vector using Invitrogen's Gateway Cloning system. Things were going smoothly for a while -- I made a pENTR entry clone, chose a pDEST destination vector, did the LR clonase reaction according to the instructions in the kit, did a diagnostic restriction digest to confirm that my plasmid had the expected sequence, and then... I couldn't maxi-prep the plasmid. This was my first time using the Gateway system, so I expected to hit some bumps in the protocol, but I've been doing maxi preps (with Qiagen kits -- regular and the HiSpeed) for years. So, I felt pretty stupid when my experiment broke down at that point. I got my plasmid out of a mini prep, but failed to recover the plasmid after inoculating a maxi-sized culture with the same exact construct in the same exact competent cells. What the hell?

As I grew increasingly frustrated, my adviser had the outlandish idea to repeat the experiment with good controls. So I did. I used every control I could think of. I transformed, mini-prepped, and maxi-prepped: my entry vector, a known entry vector with a similar sequence that we had lying around the lab, my expression vector, a similar expression vector from the lab, and the empty entry vector. I also repeated the LR clonase reaction with my plasmids and with other Gateway plasmids we had on hand.

I could get maxi prep DNA from everything except LR reaction products (i.e., any sequence of interest flipped from pENTR into pDEST through attR/attL recombination). For some reason, the expression clones would grow in mini preps, and some sort of antibiotic-resistant bacteria would still grow in my maxi cultures, but I would get abysmal plasmid DNA yield from the maxi preps. I tried growing my bacteria at a lower temperature over a longer amount of time, which sometimes helps bacteria keep larger plasmids (my plasmid of interest was about 9 Kb), but that didn't help. I bought fresh new LR clonase enzyme, but that didn't help either.

Eventually, senior lab members remembered someone having a similar problem a few years ago. They thought that the issue was finally resolved by making fresh pDEST vector by culturing bacteria containing the empty vector from the original glycerol stock. I thought that made no sense, but it was easy enough to dig the glycerol stocks out of the freezer, grow up some cells, and prep the DNA. So I did. I then used the newly made pDEST DNA to repeat the LR clonase reaction with my entry clones, transformed the LR product, and grew up some bacteria for a maxi prep.

And... it worked!

I don't know why it worked. I guess somehow, after being stored in the freezer for a while and going through enough freeze-thaw cycles, the pDEST DNA that we had stockpiled went bad. It worked well enough to take up the insert from my pENTR clones and grow in competent cells in a 2 mL culture, but then it petered out when I tried to grow a 150 mL culture of the same cells. So, if you are having a similar problem, I say to you: find your old glycerol stocks. My labmates say that the last time they had this problem, transforming the plasmid back into some ccdB-tolerant cells (the pDEST vector we used contains the ccdB suicide gene) and re-prepping from them also worked, but I can't verify that.

Next up: transfecting my expression vector into some mammalian cells to see what happens! (Also known as: the experiment that I actually care about, which I was sure I would have gotten to by now...!)