# Janet B. Matsen:Best Lab Practices

(Difference between revisions)
 Revision as of 19:24, 7 December 2012 (view source)m (→Always read the manual & don't hesistate to call tech support)← Previous diff Revision as of 19:35, 7 December 2012 (view source) (→"Did I do that?" (Bench Work))Next diff → Line 35: Line 35: #When you get your sequencing results back, you move each tube to a new box.  If the sequence confirms your construct, move each to a special "verified" box at its respective temperature.  Give each "commit" a unique number and document it in a google spreadsheet.  Cultures that are not perfect get moved to "trash" boxes that you can hold on long enough to be *sure* it is trash. #When you get your sequencing results back, you move each tube to a new box.  If the sequence confirms your construct, move each to a special "verified" box at its respective temperature.  Give each "commit" a unique number and document it in a google spreadsheet.  Cultures that are not perfect get moved to "trash" boxes that you can hold on long enough to be *sure* it is trash. #I'll make a diagram of this system. #I'll make a diagram of this system. + + == Your time is too precious to make mistakes. == + Imagine you have a simple experimental question that you expect to work for 2 weeks in order to answer.  Imagine that it turns out to take you 2 months for you to figure out that something bad happened in the beginning... "oops, the plasmid wasn't sequenced there and it is actually a problem," "oops, my strain stock wasn't pure and I'm getting mixed results," "what if these results are funny because I mixed up two tubes one day, and the mistake is propagating" etc.  This sort of devastating problem is entirely avoidable; you don't even have to be so focused that you develop a nervous twitch. + What is the solution?  Learn to recognize the moments when you '''do''' need to be hyper-alert.  Anticipate all the ways these problems can arise so you will be motivated to be hyper-alert at the appropriate times.  For example, say you want to start cultures that you will use tomorrow to inoculate for an experiment.  Any time you move cells from one source to another (freezer --> plate, plate --> overnight, overnight --> fresh media) you should: (1) move the source and destination tube/plate/whatever away from the others if others are present.  Make sure they are labeled the same thing.  As drivers look "left, right, then left again," you should grab your source, move the cells to the destination, then double check that the destination was what you expected it to be.  This is a double check that takes almost no extra time, and will reduce the chances you will make a mistake that could take you a long time to figure out. + Next, learn to experience joy when you are in a moment of extra focus!  The confidence it gives you with your work is quite a reward! =="Did I do that?" (Bench Work)== =="Did I do that?" (Bench Work)==

Back to Janet

## Always read the manual & don't hesistate to call tech support

Love thy operation manual for all equipment, and they tech support when you encounter difficulty! I am a manual-addict; I find that a good skimming of the manual for the equipment almost always give me intuition that enables better experiments and better data analysis. Tech support can be a fantastic resource when you have questions beyond the published literature. For instance, I called BioRad many times while learning to pour & run protein gels. In my quest to find out why our lab's house-made gel's buffer was wonky, I became a gel expert. Everything I learned is found here.

## Always sit down and plan at your computer before an experiment, then diligently take records

My workflow is:

1. Start a google doc in which I type up the goals of an experiment, the materials used (strains,plasmids,media) and describe the workflow. {Sample doc}
2. Detail all the information in a google spreadsheet. What are the archive numbers for the strains/plasmids? What will they be referred to within the experiment, how many biological replicates will you use, etc. {sample google speadsheet}
3. (optional): sketch up an inkscape doodle of the workflow. This is great for taking notes on as the action happens, and for understanding what you did after months/years, differentiating similar experiments, and communicating to your team.
4. Each day, plan out the nitty gritty of the experiment. Put all of the replicates and the details in a spreadsheet. Print it out the day of for note-taking.
5. If you have pictorial bits, paste them into the google doc with the date, the specifics of the picture, and the conclusions. If you are annotating protein or DNA gels, it is helpful to keep a powerpoint presentation with the pictures in time order so you can bootstrap each annotation from that you used to annotate the previous picture. This also lessens the possibility that you will forget to add key information such as the date.

Note:

• each printout should have the date (and ideally the time) at which you print it, so you can piece the papers together and not worry if you take one out of the stack for a while.
• I am very conservative about record taking, and I have never once regretted it. In fact, I always feel like I could do better than I did. The metality can be "oh, I'm just doing one thing" and it won't even occur to you to write some things down, but a one-day experiment can turn into a month-long saga. When this happens and your notes from the first few days/weeks are lacking you might end up wasting time trying to remember and recalling incorrectly.

## Efficiency Versus Thoroughness

I have really come to appreciate the trade-off between thoroughness and efficiency/throughput. If you are overly thorough and careful, things will get done too slowly. If you are overly throughput oriented, errors will occur and you will waste too much time figuring out what went wrong. I am always looking to tune my sweet spot in all of my lab practices, and always looking for tips from others.

## Record Keeping

• Page numbers are usually more useful than dates. I put page numbers (and usually dates) on all tubes I put in the freezer.
• Keep a digital list of every primer you order that you can query and reference. I number each primer tube and have them arranged chronologically in my freezer.
• I prefer to use a sketchbook rather than a lab notebook. You can buy them cheaply at art stores, they have bigger pages, and most importantly, they don't have distracting lines. This allows me to flip through pages and look for shapes (eg a table of DNA concentrations or a particular gel) more quickly.

## Committing to the freezer

• When you sequence a strain and find you created what you intended, it is time to "commit" it to the freezer. It is possible/likely that you had several incorrect constructs sequenced along the way. In general, how best can you keep organized so that you don't ever store the wrong strain, keep an incorrect plasmid, etc? I am really excited about the flow I developed for myslef:
1. Take a single colony that you will screen and inoculate two cultures with the same stick. One will be for a DMSO/glycerol stock at -80oC and one will be for miniprepping.
2. The next day when your cultures have grown up, miniprep one culture and archive the other in a DMSO stock at -80oC with the description, date, and page number on both tubes. These should both go in "pending" boxes where only un-verified constructs live; the DNA will be stored at -20oC and the cells at -80oC.
3. When you get your sequencing results back, you move each tube to a new box. If the sequence confirms your construct, move each to a special "verified" box at its respective temperature. Give each "commit" a unique number and document it in a google spreadsheet. Cultures that are not perfect get moved to "trash" boxes that you can hold on long enough to be *sure* it is trash.
4. I'll make a diagram of this system.

## Your time is too precious to make mistakes.

```    Imagine you have a simple experimental question that you expect to work for 2 weeks in order to answer.  Imagine that it turns out to take you 2 months for you to figure out that something bad happened in the beginning... "oops, the plasmid wasn't sequenced there and it is actually a problem," "oops, my strain stock wasn't pure and I'm getting mixed results," "what if these results are funny because I mixed up two tubes one day, and the mistake is propagating" etc.  This sort of devastating problem is entirely avoidable; you don't even have to be so focused that you develop a nervous twitch.
What is the solution?  Learn to recognize the moments when you do need to be hyper-alert.  Anticipate all the ways these problems can arise so you will be motivated to be hyper-alert at the appropriate times.  For example, say you want to start cultures that you will use tomorrow to inoculate for an experiment.  Any time you move cells from one source to another (freezer --> plate, plate --> overnight, overnight --> fresh media) you should: (1) move the source and destination tube/plate/whatever away from the others if others are present.  Make sure they are labeled the same thing.  As drivers look "left, right, then left again," you should grab your source, move the cells to the destination, then double check that the destination was what you expected it to be.  This is a double check that takes almost no extra time, and will reduce the chances you will make a mistake that could take you a long time to figure out.
Next, learn to experience joy when you are in a moment of extra focus!  The confidence it gives you with your work is quite a reward!
```

## "Did I do that?" (Bench Work)

• Often we do things with many tubes in parallel, such as screening colonies. One of the worst feelings is having a moment of doubt whether you pipettted into the right well in a strip of PCR tubes or whether you added a particular reagent to a master mix tube. It can happen because tubes in the middle of a row are hard to distinguish, or because a though or conversation distracts you. Tricks I have enjoyed to prevent this:
• having dogmas about the order you do things in such as "always add VF2 before VR," "always handle the tubes left to right," and "put pen dots next to reagents/steps you have added/done in a list of ingredients/steps."
• moving my eppendorf tubes from one location in the holder to another once you have taken/added material
• makes sure you don't forget one or do one twice
• drawing on PCR tubes so the wells are grouped in sets of 2-4, or by numbers relevant to how I want to order my samples.
• This gives me a point of reference when I pipette precious samples around. For example, if I am screening 7 colonies from project A, 4 from project B, 2 from C, and 8 from D, I will use a sharpie to mark the plastic after the 7th, 11th, & 13th tubes. Then as I put my samples into the PCR tubes I will know I am getting one sample in per well and have not skipped a well or put two samples in one well.
• Use a multi-pipette if you have one available and you are working with many tubes (such as PCR tubes)
• This will greatly reduce errors caused by injecting things in the wrong tube.
• write out the recipe for a master mix, and put a dot next to ingredients you have added... every time.
• I used to do this on occasion, but I recently decided to commit it to my "always" list. This is the quickest way to diagnose a failed PCR. Just recently I was *sure* I added everything but go no bands. Sure enough, I looked at my notebook and saw there was no dot next to "template." Couple this trick with "always check that each item has a dot before you use it" and you are golden!

## DNA Manipulation

• If you are doing something at all complicated, make a cad design of it before you start your DNA work. This will make you sure you have your design right, allow you to check your primers properly, and be available for consultation/sharing at a later date.

## Fun

• Science should be fun! Treat everyone well and keep your workplace a pleasant place for everyone to come.