My lab-side manner could charitably be described as ‘cautious’ – although ‘paranoid’ is probably closer to the truth. I am reluctant to take shortcuts in an experiment, for example, even when I suspect that it will make little difference. And I’m one of those annoying people who record their DNA concentrations on the tube as 1.998 micrograms per microliter instead of 2 – simply because the spectrophotometer is able to deliver readings to that level of fussiness.
Significant digits A Gilson P2 Pipetteman dialled to 0.165 microliters
This obsession extends to my pipetting. Yesterday morning, I did a few calculations and worked out that I’d need to add 0.165 microliters of an RNA oligo duplex to achieve the correct concentration. Of course I could have rounded to 0.17, because my trusty P2 is probably not that accurate, but I didn’t. (And yes, I brought the final volume up to 10 microliters with 9.835 microliters of water, though on a P20 instrument, the closest you can dial without microscopic vision is about 9.84).
I’ve always been like this. In school, I remember being affected very deeply by a math class discussion about precision versus accuracy: the classic illustration of the arrow-pierced bullseye target is still emblazoned on my long-term memory (see Wikipedia for a version of this). Even when I am cooking, I tend to measure water to the bottom of the meniscus in the measuring cup, and dry ingredients to two decimal places on our swanky digital scale. I may not always be accurate, but by God am I precise.
So is this behavior of any real value in the modern molecular biology laboratory, or is it all a sort of superstition? I guess I think that if you measure things carefully with the same instrument, all of your manipulations should be precise, if not accurate, on a relative scale. I know that there is no effective difference between 0.043 and 0.04 in practical biological terms for the experiments I’m doing, but it’s possible that my experiments are more reproducible as a result, and possibly more comparable when done on separate occasions.
I must confess, however, that I’ve caught myself relaxing a little bit on this point recently, for processes that I am pretty sure are fully robust. Only this morning I threw caution to the wind and diluted an antibody 1:100 in buffer by adding 1 microliter to 100 microliters instead of to 99. And I felt a lovely illicit little thrill in the process.
Sometimes you have to live dangerously.
It would probably have been more
accurateprecisebetteraccurate to use more volume of the RNA oligo and the water (and then use a bigger volume of the diluted volume) because I don’t think the P2 can really make a difference at that very small volume.I
dodid (that past tense will take some getting used to) always dial the pipette to the exact thing I needed, but they were not calibrated to thataccuracyprecision between different pipettes in the lab. So allis (oh, crap)was fine when I repeated things, but when someone else would give me a protocol of their dilutions, it would still be different if I then repeated that with my own pipettes.Can the P2 not make a difference at that small volume? I always figured if there is a mark for it on the dial, because the instrument is analog you must get slightly more liquid on one mark than another. Otherwise why have those marks at all? Also, when I are pulling up, say, 0.1 microliter in a fine tip, I swear I can see the different in the tiny droplet that results. Well, at least I used to be able to: my close vision has definitely started to deteriorate in my old age!
I think it’s more accurate in the middle range of the pipette than at the very low end of it. So that 1.05 vs. 1.10 makes an obvious difference, but 0.05 vs 0.10 is harder to achieve with the same pipette.
This does not have P2 in it, but the same idea: the low range of every pipette is not as great as the middle or high end.
But if it’s all relative, and you are always using the low end of the scale, could this not aid in reproducibility if you dial to the precise line rather than if you rounded up some days and down others? (I always hated the arbitrariness that said that 1.5 had to be rounded up to 2 when you had only one sig fig! It seemed so terribly unfair to the other end!)
I’m not sure. I don’t know if the problem with the low end is that it doesn’t actually measure the volume it says it does, or that it just has a lot of variation at that end. We need a pipette expert! If you have Gilson reps or pipette calibrating services come to your lab you should ask them! The lab manager in my old lab also knew this kind of thing, but that’s why I don’t – only one person needed all that knowledge about equipment!
Mind The Gap (TM) apologizes unreservedly for this geeky comment thread.
Ha. This was probably the height of intelligent discussion for 2009. We peaked too early, and now it’s all downhill from here. It’ll be sarcasm and puns until April or so, then that will slowly degrade to potty humour and knock-knock jokes over summer. By October, all comments will contain emoticons only, followed by two months of cat pictures.
Heh. And in parallel, Henry Gee’s posts will degenerate into filthy limericks.
I think you need to consider first what the impact of exactness vs. consistency will be on the outcome of the particular experiment. Then decide if the P2 (more likely to be at “its best” in the mid-range) is the right device or, as Eva said, you should use a larger volume. Now, I must admit that when I cook I look at the recipe as a guide only and am much more likely to change the proportions and even some ingredients as I see fit, so my advice may not be helpful.
Straying from the cookbook…I never do that on the first try of a new dish. It seems disrespectful not to at least give it a go as its creator intended.
Regarding larger volumes, it’s all very well if the reagents you are working with are cheap and abundant. But if they aren’t, and if they store poorly at low concentrations (as, for example, with siRNAs), then I don’t really have that luxury for low through-put experiments.
I’m going to challenge you to do the experiment one of these days…
What, stray from the cookbook?
I think a little OCD tendency is good in science, especially when working with very, very small things. But diluting 1:100 as 1 into 100? Sacrilege! Someone, quick! Arrest her!
more animals
The urge to test my precision and accuracy on those guys is extremely strong.
Now where did I leave my quiver?
The one on the right is just trying to tell you about the meniscus.
As a fellow (recovering) member of the “1:100 means 1uL stuff + 99uL buffer” club, I’m impressed that you added the extra uL with such wild abandon.
OK, last one until
Novemberafter lunch:more animals
yes, stray.
And you have an accurate balance, don’t you? See if those decimal places are reproducible.
Cath, I am immeasurably reassured that someone as sensible as you is ‘recovering’. That means there is hope for me!
Richard, you have a problem with my cooking, you can order a takeaway next time.
It wasn’t the cooking, Jenny.
It was the way it was served in 6 well plates.
Quitting bench worked helped – cold turkey.
Oh, I’d never serve Richard cold turkey in six-well plates.
Yes, I see what you mean. But I quit science for four years and it didn’t make a dent in my paranoia. Other suggestions?
so what would you serve cold turkey to me in?
It’s just a relapse. Most smokers go through several smoke-free periods before quitting for good, right? This is the same thing.
There’s nothing wrong with being a
control freak, um – precise about your work. However, your pipette will only be as accurate/precise/shiny as the tips you are using. All the witchcraft in the world won’t help your reactions if the tips are pants!True, true. I like to think my tips are above reproach, but then I would think that, wouldn’t I?
I guess it just seems weird and wrong to calculate that you need 4.77 microliters of something and then deliberately dial 4.80. If the capacity is there, why not use it?
Degenerate into filthy limericks?
Just had a quick look, and Gilson say that the random error (which is going to affect precision) on the P2 is better than 0.012 µl.
So in fact, Jenny, you could go to 3 d.p.
Dare ya.
A scientist faced with frustration
At his lack of ejaculation
Said ‘I can manage it yet
With my Gilson Pipette
And so achieve coital elation’
Dear God, what have I done.
Richard: you’ve just made my week. That news was precisely my cup of tea.
I have an audio recording of Henry reciting this limerick from memory.
It’s all LIES, I tell you.
Eva, no one will believe that until you post it somewhere…
I know, and it will be somewhere. I need to decide what to do with the rest of the conversation, and figure out audio levels so people don’t go deaf when I start laughing. It’s a conversation between 5 people in a bar, while I test out the four channels of my digital recorder. And we talk about other people, but all good stuff (well, maybe the part where we all stop believing in the existence of Neil Saunders is not necessarily “good”, but certainly interesting! )
You know, I am shocked, shocked I say that there is no protracted discussion of the correct wheel-spinning technique for dialing in such accurate measurements. I spent years dialing over volume by at least a turn and a half, then gently, gently dialing down until the right volume was displayed. 1/10th of a turn too far and you’ve blown it… dial back up and start again.
Possibly this is a completely useless exercise. It’s a bit academic now, since the last time I pipetted something, Richard Grant was still in Cambridge.
I think I need Cath’s OCD kitteh picture now.
Henry can manage it yet, with his Gilson pipette (Not safe for listening at work without headphones, or in an area in which laughter is inappropriate.)
(And some explanation and review )
It’s Henry. No explanation can ever suffice.
_You know, I am shocked, shocked I say that there is no protracted discussion of the correct wheel-spinning technique for dialing in such accurate measurements. I spent years dialing over volume by at least a turn and a half, then gently, gently dialing down until the right volume was displayed.
Richard, I am shocked that I never heard about this method! Could I be even more precise? — lord help up. And interestingly, it is very similar to the correct way to tune a stringed instrument: deliberately undershoot and then slowly bend upwards until the harmonic rings true. Fabulous.
Jennifer – I may be making this all up in my subcloning-addled brain. It was a long time ago.
Aha! The University of Alberta apparently agrees with me… PDF document right here.
Vindication, possibly.
You know, I am shocked, shocked I say that there is no protracted discussion of the correct wheel-spinning technique for dialing in such accurate measurements. I spent years dialing over volume by at least a turn and a half, then gently, gently dialing down until the right volume was displayed.
…
This method was drummed into me by a very good technical scientist during my PhD…I thought it was standard? Are there other ways?
As many as there scientists, Meagan.
I learned Richard’s wheel-spinning technique as well. It’s rote, I just picked up a pipette and checked (I couldn’t remember). Not a turn and a half, but a good half.
Jenny – I just read Experimental Heart, with great pleasure, on the eight takeoffs-and-landings-during-which-a-laptop-full-of-grant-requests-in-progress-is-off-limits I effected since last Sunday. It felt familiar – dare I write, accurate – and yet had delightful overtones of Harlequin romance and Robin Cook novels, all of which I’ve enjoyed in the past. Great fun and a perfect vindication of the LabLit principle. Thank you for writing it! (Next time I see you and when I recover my copy, I’ll ask for an author’s signature.)
Richard W – I am most shocked that you were able to find that procedural document.
Heather – my l33t 6006l3 sk1llz will not be denied!
Note that I didn’t find it from Gilson, or any other manufacturer of pipettes, which is what I was expecting. Perhaps someone can go find an instruction brochure somewhere and see if it agrees with all this.
Eight take-offs and landings? I smell a blog post… 😉
I am truly befuddled that nobody taught me this pipetting procedure. My entire world has shifted.
Heather — many thanks! Would you be willing to rate it on Amazon.com? I would really appreciate it.
Richard – perhaps, but when? By then, I’ll have put the experience behind me, like other moribund topics.
Jennifer – I’ll add that wish to the list. I’m not sure I’ve gone far enough in identifying myself to Amazon to be credible as a reviewer, but I’d do it for the principle of it…
You can even do it with a pen name — I don’t think it matters if you identify yourself; it’s all about creating a buzz. Thanks!
I had to find this old post as a “further reading” for something I just wrote and read the comment thread again, and, oh my, I think I was spot on with my prediction of how commenting would progress during the year
\o/ 😀
False precision.
I keep getting manuscripts to review with claims that p=0.0000. SPSS and physicians who are not required to learn statistics before attempting to publish them have a lot to answer for… Blood pressure measured to two decimal places is another favourite.