Home > Community > Non-linear fluorescence intensity: how to solve this ?
Upvote

19

Downvote
+ Biochemistry
Posted by
Luzze Arthur

Non-linear fluorescence intensity: how to solve this ?

Blair Blakely  Follow
Would a 4-PL fit not be more appropriate?More
Upvote

VOTE

Downvote
Gene Ewald  Follow
yes, I can fit it but I can't see, to find any references where people would use non-linear curves for fluorescence calibration.More
Upvote

VOTE

Downvote
Amir Nadeem  Follow
I would dilute/set up experiment samples to fall within the linear part of the S. This is the only range of concentrations that you're actually measuring accurately, everything above and below is out of range for the assay. Just make sure you have a good amount of data points in the linear partMore
Upvote

VOTE

Downvote
Gloria Dumaliang  Follow
What about log transformation for both concentration and fluorescence? And then fitting a linear fit to log-log transformed and then back calculating everything?More
Upvote

VOTE

Downvote
Jim Hopper  Follow
Thanks for your answer. I will try to stick to doing that. The only issue is that is linear at very low concentrations or high, so will be painful to work out correct dilutions for experimental samples, I suppose.More
Upvote

VOTE

Downvote
Chris Frost  Follow
Sounds like you're saturating. I'd just dilute down to get to the linear portion of the curve. As tempting as it may be don't use a sigmoidal model for this as that isn't the appropriate fit.More
Upvote

VOTE

Downvote
Gary V Deutschmann Sr  Follow
Yes, dilution could work in this case it's just guestimating the dilution factors for experimental fluorescence readouts that may be tricky because I don't know how all the compounds perform, for example. Let's say after 24h incubation with cells, some may reach 90% fluorescence of original concentration, and others may reach only 5%.More
Upvote

VOTE

Downvote
Charles Danley  Follow
How hard would it be to use different color fluorophores?More
Upvote

VOTE

Downvote
John Copeland  Follow
Hi. It would be straightforward as far as the chemical reaction goes, but it's expensive to buy a new dye + characterising all the compounds again would be sad! But as a last resort, I suppose!More
Upvote

VOTE

Downvote
Cars and Trucks  Follow
Is your x-axis linear or log? If it’s not log then switch it to that and see if you curve is linear nowMore
Upvote

VOTE

Downvote
John Miller  Follow
Thanks. But is Log/Log OK to use for fluorescence ? It's linear-ish with log/log both axisMore
Upvote

VOTE

Downvote
Dave Hunter  Follow
It sound like at higher concentrations your compounds are either insoluble, aggregating, or undergoing some other form of quenching. You should work within the linear range of your compounds. The problem with fitting to a non-linear is that you are essentially incorporating a fudge factor that may not fit for every scenario, as it may not be actually reflective of probe photophysics but a context-dependent phenomenon like those listed above. In terms of troubleshooting on a per compound basis, you could check the computationally derived partition coefficients (LogD or otherwise) as a starting point, this should give you an idea of lipophilicity. There are also some additionally assays that can be run (DLS, shake-flask LogD, etc) that will help you gauge linearity but it sounds like your standard curve itself is giving you a good idea of that. In terms of the different intensities, so you mean they show different intensities at the same concentration? Even low concentrations? This could be indicative of some type of intramolecular quenching such as PET. In my experience some of these probes will “turn-on” when they enter a cell and this could really throw off results. Typically we adjust linker length and fluorophore structure to account for this.More
Upvote

VOTE

Downvote
Ian Yates  Follow
Again, many thanks for replying here!More
Upvote

VOTE

Downvote
Food Lovers - Spot  Follow
And here is the full calibration curve of 4 compounds: https://postimg.cc/gr5pMX6pMore
Upvote

VOTE

Downvote
James Freericks  Follow
I suppose to account for differences within linear ranges between compounds, I could calculate the correction factor, for example, slope/average sloped, then multiply each fluorescence intensity by the corresponding correction factor.More
Upvote

VOTE

Downvote
Jim Casey  Follow
Just to have an idea:More
Upvote

VOTE

Downvote
Gaye Guida-Dennis  Follow
Thanks a lot for this reply; super useful! I will give it a go at logD/logP. At lower concentrations, it's not too different, but there is some difference still.More
Upvote

VOTE

Downvote
more replies