This package assigns single-molecule identities based on the salvaged fluorescence approach.
- 1.0
- Microsoft Windows 7 or newer, 64-bit
- Matlab R2017b or newer
- Curve Fitting Toolbox
- Optimization Toolbox
- Download and install Matlab of the right version (30-60 min)
- Download the code to your computer (a few seconds)
- Start Matlab and run the code
example1.m: this example shows how to estimate the photon number in the salvaged fluorescence channel.
example2.m: this example shows how to perform color assignment based on the photon numbers in the conventional and salvaged fluorescence channel.
Conventional_fitting.m: this script shows how to perform single-molecule fitting in the conventional fluorescence channel and estimate the positions and photon numbers. It saves the fitting result as "data_example1.mat", which will be loaded in example1.m. If the fitting result does not exist, it will run this script automatically to generate the data.
The optional step uses a GPU single-molecule fitter from "fit3Dcspline"(https://github.com/jries/fit3Dcspline; https://www.nature.com/articles/nmeth.4661)
The GPU fitter requires:
- Microsoft Windows 7 or newer, 64-bit
- CUDA capable graphics card, minimum Compute Capability 3.0
- CUDA 8 compatible graphics driver (for GeForce products 378.66 or later)
An example two-color dataset of a COS-7 cell imaged with the salvaged fluorescence approach.
Download the dataset from: https://www.dropbox.com/sh/qw18jf3qcma1pbj/AACVAfhZ0WxM87vW-5VgrvIaa?dl=0
- Raw images (3000 frames):
- Conventional_fluorescence.tif: images in the conventional fluorescence channel, acuiqred by a sCMOS camara
- Salvaged_fluorescence.tif: images in the salvaged fluorescence channel, acuiqred by an EMCCD camara
- Fitting results:
- data_example1.mat: this is the fitting result of "conventional fluorescence.tif", generated by running "Conventional_fitting.m"
- data_example2.mat: this is the analyzed result (provided by the authors) of the full dataset (150,000 frames) that used in example2.m to perform color assignment and reconstruction
Additonal information of the dataset
- Labeling:
- Channel 1: overexpressed GFP-Sec61b labeled with anti-GFP primary antibody and CF660C conjugated secondary antibody
- Chennel 2: anti-a-tubulin primary antibody and AF647 conjugated secondary antibody
- Imaging conditions:
- 200 fps
- 642 nm laser at 15 kW/cm2
- only the lower objective was used to collect the fluorescence
- Microsoft Windows 10 64-bit
- Matlab 2017b
- CPU: Intel Core i7-6850K
- GPU: NVIDIA GeForce GTX 1080 Ti (Driver version: 391.01)
- CUDA v4.2
- Conventional_fitting.m: 35 seconds
- example1.m: 11 seconds
- example2.m: 115 seconds
- Conventional_fitting.m: this script saves the fitting result as "data_example1.mat"
- example1.m: this script will display a scatter plot showing the intensity of of single molecules in the conventional and ssalvaged fluorescence channels. The plot is similar to Fig. 1D in the manuscript.
- example2.m: this script will display a bined 2D intensity histogram of AF647 and CF660C (similar to Supplementary Fig. S4G) and the overlaid image of ER and microtubules (the same data shown in Supplementary Fig. S13D).
- Save the images in the same format as the demo dataset
- Change the file names accodringly in the Matlab code
- It requires a resgistration file between the two channels by imaging fluorescent beads
- Follow above instructions to run the code
For any questions / comments about this software, please contact Bewersdorf Lab.
Copyright (c) 2019 Bewersdorf Lab, Yale Univeristy School of Medcine, USA.
The package is licenced under the GNU GPL.
If you use SF_color_assignment to process your data, please, cite our paper:
- Zhang, Y. et al. Nanoscale subcellular architecture revealed by multicolor 3D salvaged fluorescence imaging. bioRxiv, 613174 (2019).