Team:Bielefeld-CeBiTec/Design

iGEM Bielefeld 2015


Output Signal processing

A prototype for easy fluorescence detection

The problem

The aim of our project was to design a convenient test strip using fluorescence as an output signal. The green fluorescent protein (GFP) is one of the most frequently used reporter proteins and it has already been used in numerous iGEM projects, e.g. iGEM Aachen 2014. Instruments for fluorescence-measurement are not readily available outside of the lab. So our aim was to develop an easy and cheap method for analyzing a fluorescence output. Therefore we developed an easy and cheap method for analyzing a fluorescence output. So we developed a smartphone app in combination with two light filters. One filter is used to ensure the correct extinction of GFP by putting it in front of the flash, while the other filter passes only the evaluable emission spectrum of GFP for minimizing the background noise. Thereby it is put in front of the camera of the smartphone.

The first steps

When starting with our project, we wanted to verify if fluorescence of sfGFP can be measured by a smartphone. So in the first step we measured the extinction and emission spectra from sfGFP to test if the smartphone flash can excites sfGFP. Although the characteristic curve of the flash of the smartphone s5 mini, do not covers the hole sfGFP exciting curve, it is possible to extinct sfGFP with the flash, because the maximum turning point is covered from the characteristic curve.

Next we tested some filters from the light engineering of the women cultural center Bielefeld e.V. ("Frauenkulturzentrum Bielefeld e.V."). We tested these filters in front of either the camera or the flashlight by taking picture of purified GFP, sfGFP lysate and lysate without sfGFP (for generating the lysates see here). Different hues and different levels of brightness can be identified in the picture below. With this first impression, we decided to test different filters. The goal was to find one filter or filtercombination which was perfectly suited for sfGFP imaging, cutting out as much of the background signal as possible.

Emission and extinction spectra of sfGFP. The characteristic curve from the samsung galaxy S5 mini flash is also shown. The overlap of the different spectra shows that it is possible to excite sfGFP with a smartphone flash.
First filter test. Purified GFP, sfGFP lysate and lysate non-induced sfGFP culture were photographed through light filters to assess the feasibility of fluorescence detection using a smartphone.

Therefore we bought a lee color filter catalog, so we can test various filter combinations in the next step. A preselection of the filters was possible, because we had access to the light transmitting spectra of almost every filter. Apart we designed a handcraft black carton for taking photos under defined conditions and a constant dark background.

Lee color filter catalog
The black box: tinkered to take photos under same conditions

The filter combinations

The essential requirements for taking comparable photos are achieved. Now we had to find the ideal filter combination. So we tested every combination of 5 filters for emission with 12 filters for extinction. A lot of pictures had much background or no fluorescence was seen. All together we tested 60 combinations to find the optimal filter combination for sfGFP imaging via smartphone with less background as possible. Afterwards we had to analyze the pictures. Hence we analyzed the pictures with the imaging procesing software Fiji. We measured all fluorescence of every reaction tube in every picture we took. After analyzing the photos with the image processing software Fiji, the optimal filter combination was found Tokyo blue (071) in front of the flash and twickenham green (736) in front of the camera provided the best result for sfGFP.

In the picture below you can see the test of different filter combinations to image sfGFP. On the top the picture was taken without any filters. In the middle the picture was taken with fluorescence green (219) in front of the camera and deep blue (120) in front of the flash. The bottom photo was taken with the optimal filter combination of tokyo blue (71) in front of the flash and twickenham green (736) in front of the camera.

Test of different filtercombinations. On the top there is the picture without any filters. In the middle the picture was taken with fluorescence green (219) in front of the camera and deep blue (120) in front of the flash. The bottom photo was taken with the optimal filter combination of tokyo blue (71) in front of the flash and twickenham green (736) in front of the camera.
Test of different filtercombinations to photograph mRFP. On the top the picture is taken without any filters. The picture in the middle was taken with light red in front of the camera and dark yellow green in front of the flash. The bottom photo was taken with the optimal filter combination twickenham green in front of the flash and light red in front of the camera.

As you can see in the pictures above it is really important to choose the right filters to get a high fluorescence signal and a low background signal. But is it only possible to photograph sfGFP and GFP? To find out, we took photos of the monomeric red fluorescent protein (mRFP) lysate as well. We tested the combination of 4 different filters for extinction with 8 filters for emission. All in all we tested 32 filter combinations to find the optimal filter combination. After this we analyzed again the fluorescence of the different reaction tubes in the pictures with the image processing software Fiji. The best filter combination for imaging mRFP with a smartphone are Twickenham green (736) in front of the flash and light red (182) in front of the camera.

In the picture you can see the test of different filtercombinations to photograph mRFP. On the top the picture is taken without any filters. The picture in the middle was taken with light red in front of the camera and dark yellow green in front of the flash. The bottom photo was taken with the optimal filter combination twickenham green in front of the flash and light red in front of the camera.

So probably it's possible to detect most of the fluorescent proteins, you just need the right filter combination and a flash which can excite the relevant protein.

Does fluorescence photography work on paper?

We showed that fluorescence imaging of liquids in quite high amounts is possible with our approach. But does it also work with small volumes and on paper? To assess this we took normal laboratory filter papers. (MN 827 B from Macherey and Nagel, C 350 L and FN3 from Munktell and a laboratory filter paper from Merck), put 5 µL of the lysates on the paper and took photos with our filter systems. Exemplary the results of the laboratory filter paper from Merck are shown in the figure below. The other filter papers showed the same results. In the first row dilution series of purified GFP were added to the paper discs. In the second row sfGFP lysate were added, in the third row mRFP lysate was used and in the last row lysat without sfGFP and mRFP were added to the paper discs. The top photo was taken without any filter. As you can see, there is no fluorescence visible. The picture in the middle was taken with the filter combination for sfGFP. In this picture you can see two glowing paper discs in the first and in the second row. So fluorescence detection of 5 µL GFP on paper works up to a concentration of 17,5 mmol/L.In the bottom picture the mRFP filters were used. Three paper discs of the third row are red. So it's possible to detect the mRFP lysate up to a 1:10 dilution.

We added 5 µL of lysate to the paper disc and photographed it with the different filter combinations. In the first row dilution series of purified GFP were added to the paper discs. In the second row sfGFP lysate were added, in the third row mRFP lysate was used and in the last row lysat without sfGFP and mRFP were added to the paper discs. The top photo was taken without any filter. As you can see, there is no fluorescence visible. The picture in the middle was taken with the filter combination for sfGFP. In this picture you can see two glowing paper discs in the first and in the second row. In the bottom picture the mRFP filters were used. Three paper discs of the third row are red. So it's possible to detect the mRFP lysate up to a 1:10 dilution.



Now we wanted to find out if it is possible to photograph the fluorescence from a Cell-free Protein Synthesis (CFPS) on Paper . After a CFPS run on paper, we took the paper discs and took the photo shown below. We took the paper discs of paper from the experiment on paper, took a photo and it worked, hence we can take a photo of the sfGFP generated in a CFPS experiment. In the first row you can see purified GFP, a dilution of the GFP and sfGFP lysat. We used this row for comparison. In the second row is the negativ control. We need the negativ control to visualize the self-fluorescence of the cell extract. In the third row are the paper discs where our self made extract were added for the CFPS. In the last row you can see the paper disc with the commercial cell extract for CFPS. So it's possible to take an image of sfGFP produced during a CFPS reaction with an smartphone. Hence it's possible to take this as a device to analyze our paper based biosensor.


Photo of the CFPS run on paper. We took the paper discs of paper from the experiment on paper, took a photo and it worked, hence we can take a photo of the sfGFP generated in a CFPS experiment. In the first row you can see purified GFP, a dilution of the GFP and sfGFP lysat. We used this row for comparison. In the second row is the negativ control. We need the negativ control to visualize the self-fluorescence of the cell extract. In the third row are the paper discs where our self made extract were added for the CFPS. In the last row you can see the paper disc with the commercial cell extract for CFPS.

The app

Now it's possible to photograph fluorescence, but sometimes it's difficult to see the differences between the divers paper discs. Additionally analyzing the pictures with the image processing software Fiji requires know-how of the user. For both the program and the interpretation. We solved this problem with a smartphone app. We coded the app with Android Studio 1.2.2 and it can be installed on smartphones on Android 4.2 (Jelly bean) and later. The app determines the median of the greenvalue of a area of pixels in the sensorspot and compares it to the pixels of the negative control (the not induced sensor). If the ratio of a sample and the negative control is above a certain theshold (10%), a valid biosensor signal is detected. Additionally it is checked, if the value for the positive control (drop of sfGFP) is above a certain value, confirming the functionality of the biosensor. Furthermore the app displays specific information regarding the different heavymetals and date rape drugs to inform the user. You can download it by clicking on the button below. You need to unzip the file on your phone, since .apk files can not be uploaded. Alternatively download it here and delete the fileextension ".txt".

We designed and printed a black case for an easy handling and a correct positioning of the smartphone and the biosensor test strip.

A photo of our test strip. The photo was generated by a typical smartphone (samsung galaxy s5 mini) using an ideal filter combination to detect fluorescence of sfGFP.
Screenshot of the app. It shows the pixels who are taken into account for the calculation in red, so the user can check, whether or not the pixels chosen are correct. Furthermore, it provides the values that are calculated for the greeness of the spot.
Combination of the app and the black case. The black case provides the perfect environment for capturing pictures under reproducible conditions. The application will calculate whether a contamination is present and give out a list of contaminants in the water sample.

The Black case

We showed that, the fluorescence detection works fine, but it's not very practical to take the photo. The filters have to be in the exact right position directly in front of the camera and the flash. It also has to be quite dark for high quality pictures. Therefore we designed a black case and realized it with a 3D printer. Marco Radukic programmed the file for the 3d print in exact accordance with our guideline. The top smartphone inlay can be specifically changed for different smartphone types. The test stripe can be placed on the push loading drawer and inserted into the box. Thus it is quite easy to use. The smartphone is put on the top inlay, the test strip is inserted and the taken photo provides information about the contamination. To download the model for the 3D printing click here.


You can see the 3D modell of the case, realized by Macro Radukic in exact accordance with our guideline. On the right you can see the printed black case. Printed from Thomas Schäfer with his 3D printer.