[FLIMfit-users] Dundee FLIM data

[Munro, Ian]

Hi Markus

Firstly I hope it’s ok if I cc this to the FLIMfit-users mailing list
lists.openmicroscopy.org.uk/mailman/listinfo/flimfit-users<http://lists.openmicroscopy.org.uk/mailman/listinfo/flimfit-users>

just in case anyone might have similar questions.


An IRF gives the software 2 important pieces of information.
It allows it to estimate the ‘distortion’ introduced   by the instrument and ,crucially,  to estimate the position,  in time , of the excitation pulse.

In the absence of an irf,  FLIMfit assumes a “delta function-irf” which i.e. a perfect detector.
The Excitation pulse is then assumed to be at the first non-zero point in your data (after all, given a perfect detector, no fluorescence can be detected before the  excitation pulse).

In the absence of a measured irf you can use this to do a “tail-fit” .
To do this, estimate the peak of the decay by eye & exclude all earlier points from the decay using the time-min box on the data tab.

the attached screen shots illustrate this.

Beware though.  A tail-fit is fine for qualitative analysis. ( does  area a) have a shorter lifetime than area b)   ).
For quantitative analysis it is very dangerous as you are entering a crucial parameter ( the excitation pulse position ) by ‘eye’.

To select a region use the shape tools (square polygon & ellipse shapes) at top left.

Regards

Ian

[cid:CEFAA302-69A9-457D-B02E-A99762A13589]



Exclude data before peak.

[cid:FB65B70A-8BFF-4FEC-BE10-56A79F3EBC81]


Fit !   Note red line & residuals.

[cid:178D0F88-6EB9-4809-B03B-3F09A229BFB2]


On 12 Sep 2014, at 14:53, Markus Posch <m.posch at dundee.ac.uk<mailto:m.posch at dundee.ac.uk>> wrote:

Hi Ian-

thanks for your help in the first place!
I managed to get hold of the correct TCSPC data - but still much lost in the data analysis.

Particularly I am not sure if I am missing out an IRF (instrument response function).

We think want to measure the half-lives with small areas - so basically within an ROI. Is it possible to define an ROI and get the T-half within this ROI?

For the moment we having two images only (control and FRET cell) for a brief proof of concept.
Eventually and for statistical reason we’d have to do may more. So potentially displaying those in a histogram of distribution of lifetimes in all ROIs measured.

I have attached brief figure how we think we wanted to represent data (numbers in table are made up).

Is such analysis possible? One could do that on the LaVison software at the Beatson, if I remember correctly.
Could you please guide me through a short workflow?

Data I have are two OME-tifs (of the TSCPC recording) - IRF needed?

Many thanks,

Markus


The University of Dundee is a registered Scottish Charity, No: SC015096
<TCSPC.pdf>
<FLIM-FRET_histo.pdf>


On 12 Sep 2014, at 12:14, Munro, Ian <i.munro at imperial.ac.uk<mailto:i.munro at imperial.ac.uk>> wrote:

Hello Marcus

I was concerned that FLIMfit wasn’t able to handle your data but from what you say,  perhaps the first step is to check that it actually is time-resolved. We do have some FLIMfit  users at Beatson already so perhaps they may be able to help.

Unfortunatley FLIMfit is principally  a time-domain tool in terms of fitting.
It might, however, be of use in terms of visual examination of the data.

Colleagues have looked at converting data between time-domain & frequency-domain (FFT I think) in order to compare the two approaches but I haven’t been at all involved.
I could put you in touch if that would help.

All the best

Ian




On 12 Sep 2014, at 10:38, Markus Posch <m.posch at dundee.ac.uk<mailto:m.posch at dundee.ac.uk>> wrote:

Hello Ian-

thanks for your mail and may thanks for offering your help!

We did some preliminary FLIM experiments at the Beatson Institute and are trying to get a feel for FLIMfit on these data. Early days, so we still have a steep learning curve and all needs to make slowly sense to us.
Data is from the LaVison TCSPC and the Lambert Inst. LIFA system.

We have tried to upload some TCSPC data into FLIMfit, but it appears this was only RGB-Tif snapshots, rather than the actual dataset.
Petr has identified there was clearly no time dimension in the data. - I will try to check with Beatson on the data!

I could also upload the data (this was just a pilot with a handful of cells) on OMERO. What server would make sense?

Your offer to help is certainly most welcome!
Will FLIMfit work with TSCPC data only or also process frequency modulated data, as from the LIFA system?

Many thanks!

Markus

The University of Dundee is a registered Scottish Charity, No: SC015096



-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openmicroscopy.org.uk/pipermail/flimfit-users/attachments/20140912/0159ddfb/attachment-0001.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Screen Shot 2014-09-12 at 15.38.31.png
Type: image/png
Size: 214385 bytes
Desc: Screen Shot 2014-09-12 at 15.38.31.png
URL: <http://lists.openmicroscopy.org.uk/pipermail/flimfit-users/attachments/20140912/0159ddfb/attachment-0003.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Screen Shot 2014-09-12 at 15.39.55.png
Type: image/png
Size: 205137 bytes
Desc: Screen Shot 2014-09-12 at 15.39.55.png
URL: <http://lists.openmicroscopy.org.uk/pipermail/flimfit-users/attachments/20140912/0159ddfb/attachment-0004.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Screen Shot 2014-09-12 at 15.41.43.png
Type: image/png
Size: 208806 bytes
Desc: Screen Shot 2014-09-12 at 15.41.43.png
URL: <http://lists.openmicroscopy.org.uk/pipermail/flimfit-users/attachments/20140912/0159ddfb/attachment-0005.png>