The Colony-Picture Project

One of our goals was to build a reasonable colony picking robot to outsource the tiring task of colony picking by hand. Automated lab processes have improved in recent years and are being improved today and in the future. We used an Opentron and we constructed modules and scripts, that allow other iGEM Teams to turn their own OT-2 into a colony picking robot as well.

To achieve this goal, we initiated this Collaboration because we needed to train the artificial intelligence with a lot of data, so that it is able to recognize the colonies by itself. We asked other iGEM Teams if they could provide us with pictures of their agar plates with E. Coli colonies with which we could train the robot. In order to do that we published a script, code and module we created.

To participate all that was needed was 9 cm diameter agar plates with E. Coli colonies, a 12- inch (or larger) screen and a 12-megapixel camera (phone or SLR). Other criteria were evenly distributed, easily distinguishable colonies that don’t clump together and using LB medium only. The first step was taking the picture. For this purpose, the screen had to be set to the highest level of brightness and display a white image. Then the transparent paper we sent the teams had to be placed on top of the monitor and on top of that the agar plate was placed and the lid removed. Subsequently, the picture was taken from a 90 degree angle whereby light reflections should be avoided.

Following this, the images taken had to be labeled using a web-based tool. The picture had to be converted into a jpeg file and then uploaded in the weblink. In the weblink the option ‘rectangle’ had to be chosen. Afterwards, the single colonies were marked. In the end the pictures were saved and exported as a json-file and then uploaded in a folder in the google drive.

Next, we fed the robot the data.

As an incentive to participate in this Collaboration we gave away prizes. The first ten teams that send us 100 pictures got a small surprise, an E. Coli soft toy. The team that sent us the most pictures got a bigger surprise, a ….. @Team.