?

Log in

No account? Create an account
Lab Rats [entries|archive|friends|userinfo]
Lab Rats

[ userinfo | livejournal userinfo ]
[ archive | journal archive ]

ELISAs [Nov. 1st, 2008|10:55 pm]
Lab Rats

lab_gripes

[bubbleteagirl02]
So, first off, 'ello all!

Second, if anyone works with mice, perhaps you can help me with this question.

Do you know of any company (not R&D or PBL) that makes a good Interferon-alpha ELISA for mouse serum that is really sensitive? I just ran the R&D ELISA, know I did everything perfectly, and it stunk, royally. Thus I am frustrated at the lack of good ELISA readings (the standard curve wasn't even graphable!) and wondering if there is another good ELISA out there or if I need to start looking for other readouts for Interferon-alpha.
linkReply

Comments:
[User Picture]From: ignis
2008-11-02 03:08 am (UTC)
Unless you can get the standard curve to work I'd be more inclined to think that the problem isn't with the antibody but with your technique. That isn't a personal attack, that's just playing the probability. A quick look at your journal tells me you just graduated and are doing research for a year while you are waiting for medical school, nothing wrong with that!

1) Have the humility to know that even when you think you've done something perfectly, research doesn't work like that. That's why it is an art, not just a science. That's also why some people hate research "I did it exactly the same way as before, why didn't it work this time!?!?!" That's science.
2) Why did you use the kit/antibody you used in the first place? Have you gotten prior results that are positive? Has anyone else gotten prior results that worked? How old is the kit, is it expired?
3) Re-read the manual/instructions and pay extra close instructions about getting poor results.
4) Trouble shoot!
(Reply) (Thread)
[User Picture]From: bubbleteagirl02
2008-11-02 03:47 am (UTC)
1) My boss looked at my results and we basically came to the conclusion that the lower half of the standard curve stunk for whatever reason but a fair number of my duplicate sample wells were consistent, just lower than even the standard curve should have read, so we can't actually say anything about that. While I am totally willing to admit technique error, I was ridiculously meticulous about this ELISA since I had messed up my last one on IL-10.

2) A lab mate wanted me to use this ELISA to measure INF-a levels in mouse serum in his experiment. We had hoped that post transplant that there may have been a spike in the levels. We have never used this kit before but it is one of the only IFN-a kits on the market (unfortunately not designed for serum but we trouble shot that). Kit was not expired, came in that day.

3) The directions say nothing about poor results. I barely had enough directions to figure out how/of to dilute the samples even if I was using culture media and not mouse serum.

4) Did that the best I could. Everything was in duplicate and I even spiked extra serum samples to check for masking/depletion of a known amount of INF-a. Problem was that I couldn't get the curve and thus only have super low OD values.


(Reply) (Parent) (Thread)
[User Picture]From: ignis
2008-11-02 03:54 am (UTC)
The duplicates will tell you if your pipetting and your washes are good...

So the top half your your std curve, did you get good numbers from that? Could the issue be in the dilutions? What was the range of your curve that was attempted and what was the range that worked? Did you dilute your standards and your samples in the appropriate buffer?

If nothing else, call the company monday morning?
(Reply) (Parent) (Thread)
[User Picture]From: bubbleteagirl02
2008-11-02 04:08 am (UTC)
Sorry if I got defensive, I am really pissed at this ELISA.

Some of the duplicates were perfect, and some were weird. And it had no correlation to the samples groups or the standard, it was just... weird.

The top half of my standard curve would have gotten great readings if any of the samples had the same level as the top half of the curve. Unfortunately all my samples fell in the bottom bad half of my curve. And further, since the program wouldn't even make a standard curve, I have no idea the actual levels of IFN-a in my samples beyond my OD.

Don't know if I mentioned this, but it was a precoated plate sandwich ELISA.

The curves range should have been from 12.5 pg/mL to 500 pg/mL. Ranged that worked was the top three or four which I think was 500 to 100 pg/mL 50 if the fourth standard was good (which I can't remember if it was or not).

Samples and standard were diluted in the buffer provided by the kit. It was the first time I used a plate washer, but I did use the provided wash solution in the plate washer.
(Reply) (Parent) (Thread)
[User Picture]From: lickis
2008-11-03 02:10 am (UTC)
Sometimes ELISA is like that. The standards never actually make a straight line, it really is a curve. When you're at the low end the values can sometimes appear to be all over the place. If you're using a kit you tend to assume that they wouldn't tell you to do standards that are below the linear part of the curve because they'd be useless, but the thing is that especially when it depends on room temp for reactions the colour development isn't the same every time. So perhaps you needed to leave the development longer or something.
However, if that is the case it's basically irrelevant if your samples are at the same point in the curve. To get good results you would then need to dilute your samples less so they end up nearer the top half of the curve where it works nicely. That can be a total pain in the butt with mouse serum because you don't have of it (we need special mutant mice that have huge volumes of blood!). It can sometimes be helpful to half volumes and do a 50uL ELISA (assuming you usually use 100ul) so that you don't need as much of it. This has always worked well for me :)
(Reply) (Parent) (Thread)
[User Picture]From: lickis
2008-11-03 02:14 am (UTC)
...you don't have a lot of it, that is. I should start proof-reading _before_ posting ;)
(Reply) (Parent) (Thread)
[User Picture]From: bubbleteagirl02
2008-11-02 03:51 am (UTC)
Oh, and the last ELISA I did, I had a beautiful standard curve even though the rest of the ELISA was less than perfect (the hazards of poor sample labeling).

And just as a heads up, I may have just graduated, but I did over a year of research in college in addition to all our very hands on labs where some of our work went towards helping professor's projects. No offense but I am pretty good in lab, not perfect, but I typically can get standard curves to work.
(Reply) (Parent) (Thread)
[User Picture]From: ignis
2008-11-02 04:11 am (UTC)
Like I said, I didn't mean to criticize you. I've found research to be very zen like. When you just start out, you know you know nothing... then after a couple years, you think you know what you are doing.... then after several more years, you realize that while you know a lot, you know nothing. I spent 5 years in a lab before starting medical school and I trained many people. I got a kick out of some people's frustrations at how an experiment works one week, but not the next, and then magically started working again. It's funny how scientific and rational people can become so superstitious. The MDs had the hardest time with the rhythm of science, some times stuff just doesn't work and it's only after you prove that it doesn't work... does it really not work! Often it's because they were rushed in the lab, or their technique was off because they were so 'perfect' that they over washed something. Anyways, bed time.
(Reply) (Parent) (Thread)
[User Picture]From: bubbleteagirl02
2008-11-02 04:15 am (UTC)
LOL! Thankfully I have a low stress environment in lab, well at least when we are not rushing to get downstairs to inject the mice before the lights turn off. It was just annoying because I was very careful to pipet everything correctly and mix everything as the instructions said, even using a plate washer this time instead of a wash bottle like the direction's said and it still fails.

Anyways, thank you very much for you input! Its actually helped more than you think because I had to go back and think about exactly what I did. Still don't know why the ELISA didn't work.
(Reply) (Parent) (Thread)
[User Picture]From: kevisannasdad
2008-11-03 02:41 pm (UTC)
Don't know if you are still thinking about what might have gone wrong, but it sounds to me like you just needed to leave the color development stage just a bit longer. I bet you followed SOP on that stage but sometimes if you wash a bit more thoroughly or abbreviate an incubation, even slightly, it reduces your sensitivity. Also, because ELISAs are so sensitive, it is a good idea to make sure your pipets are well calibrated.
(Reply) (Thread)
[User Picture]From: lyssabits
2008-11-04 01:04 am (UTC)
R&D's website is a bit irritating, as they do not have anything beyond the data sheet for the assay. A couple of questions/thoughts..

kevisannasdad's suggestion to try letting the color develop longer was a good one. If it's unusually cold that could be slowing your development. If all the readings for the stands tend to be low (like, say, if your highest standard isn't much above 1) then letting the color develop longer may help. I know I've screwed up ELISAs through no fault of my own when the heater in here was broken and it was about 90 degrees. I'd also try diluting the samples less. I know mouse serum is precious, if you need it for something else it can be tempting to try and stretch it.

How are you storing the serum? How many times has this serum been freeze/thawed? Serum's finicky that way, it's not recommended to freeze/thaw it more than 3 times.

When you say your standard curve isn't graphable, what do you mean? Most of the time these curves aren't completely linear, but look more like an S, so it's difficult to detect changes at the lower and upper ends of the curve. However if they're erratic, then that's more likely to indicate a dilution error.

This kit is optimized for media and you're using it for serum? How did you trouble shoot this? (I admit, I have no idea how you'd even begin to trouble shoot for this. ;) ) My first thought was that there was something in the serum inferring with the primary antibody. Do you have any idea what the circulating levels of INF-alpha should be? Some factors circulate at very low levels and the serum needs to be concentrated.

Also noticed that you said you were using a plate washer instead of manually washing.. that does change how many times you're "supposed" to wash. Generally you can wash fewer times with a plate washer than when you do it by hand.. so be careful you're not over-washing.
(Reply) (Thread)