1. I uploaded my data but the calculation crashes / I get HTTP status 550
Our old "matlab" site was particularly sensitive with this issue. The new python based system is not as sensitive, however, you must still ensure that the input file is correctly formatted. While you don't have to (but it doesn't hurt) follow every single detail in this old video on how to prepare the excel input file that we created for our matlab-based website (see link here) you still need to ensure that the first row contains your "headers" and that column A = host concentrations, column B = guest concentration and columns C - XX (depending on how many columns of data you have).
The other possible reason is that the initial guess was very far off target or the data fits really poorly to the model you are trying to use, e.g. because the binding is exceptionally strong. Please don’t hesitate to contact us for help for assistance.
2. The calculators doesn’t show on any of the webpages.
This might be because our backend server is down. Please contact us immediately here.
3. I get a "Fit failed" (orange colour) with my input file.
This is probably because your initial guess is too far off for the fitter to find a solution. Try changing the initial guess - usually making it 10-100 lower is a good start, especially if you suspect the binding is weak (try even as low as 10 or 1). For 1:2 and 2:1 thinks might trickier again. Try "low guesses", try high "guesses", try negative and try positive cooperativity. For the 1:2 or 2:1, if nothing else works, fit it first to 1:1 - it might be a bad fit but the Ka could give you a clue towards what to put in as your iniital K1 guess for the 1:2 (or 2:1) fit. If all fails, you are welcome to contact us here.
4. I didn't save my permanent "url" anywhere. How can I get it back?
IF you put in your email address or made your data public (or both) then it is straigtforward.
i) Go to Bindfit and select "Search" at the top (shortcut here).
ii) With an email address, put your email address in the top right corner and select "Retrieve fits". Within a few minutes you should get an email (check your junk mailbox/setttings) from the BindFit database (noreply at opendatafit.org) with a list of your fits (unique URL's).
iii) If you made your fit "searchable" (public) you can search for any terms you may have included in your entry in the top left search box and hit "search". This will return list of all the fits (unique URL's) associated with your search term.
If you did NOT put in your email address or made your entry searchable then it is "almost" impossible to retrive these. Please contact us if you have to retrieve the data and you can't just re-run the calculations.
5. What is the error on my "K" values?
We have now included error estimation on our binding constants. This is the so-called "asymptotic error" at the 95% Confidence Interval level. Please see P. Thordarson, Chem. Soc. Rev., 2011, 40, 1305-1323 – Link for further information about the asymptotic error and some of its limitation. The method used in our program uses the approached detailed by R. de Levie, J. Chem. Edu., 1999, 76, 1594-1598 - see link here. We also intent to include a Monte Carlo based estimation of this uncertainty in the near future.
6. Can I save my data without doing any fits (and why would I want to do that)?
Yes you can! In BindFit simply select the "save data only" options (NMR or UV) towards the bottom. You do need to make sure your input data is formatted correctly (first raw headers, column 1 - host conc, column 2 - guest con and column 3-end measured values). Once you uploaded you data you get straight to the "save" option for the database and you get the opportunity to add additional information (meta data) to your input data before saving it to our database and getting an unique URL's.
Why would anyone want to do this you may ask? We had severval requests for this option as it allows people to archive their data using our database even if they haven't used our website to fit the data. This might include historial data that had already been proceesed with other software packages or cases where you had to use a different program / equilibria that is currently not available on our site. The benefit for you in that situation is that you can still submit unique URL's with your datasets with your publications so that everyone has access to your raw data after publication.
7. The program on this website does not have the option to do "X" (insert your pet option here). Will it do so in the future?
In most cases the answer is "yes"! But it is matter of time and resources. Please email us your suggestions. The current list of features that have been raised and we would like to include:
- Indicator-displacement model. If you got good "test data" to send this would help us getting ready.
- Dimerization model (we will hopefully have this ready very soon)
- Aggregation model(s) (we will hopefully have the first one ready very soon)
- 2:1->1:1->1:2 equilibria (working on it - good test data would help)
- Including the initial value (chemical shift or UV absorbance) as one of the parameter to be fitted (n.b. this will always create a better fit but that may not always be justified). T
- Include non-negative constraints on UV data fitting (currently the program can return negative absolute molar absorptitivies)
- Stoddart´s dilution approach (J. Org. Chem., 2001, 66, 3559-3563 - link).
- Include activities (n.b. this is HARD).
Got a question – please email us here.