¹The Rockefeller University, New York, NY 10065, USA.
²Singer Instruments, Roadwater, Somerset, TA23 0RE, England
Picking accuracy, or the ability to pick a microbial colony precisely without touching other unwanted materials, is by far the most important reason why we replace humans with robots for picking hundreds, if not thousands, of colonies in the lab. Robotic pickers provide consistency, repeatability, and are able to work for long hours.
However, few robotic pickers are able to pick colonies with high accuracy. Many robotic pickers fail to pick with high accuracy for various reasons. That is why it is important to thoroughly test the picking accuracy of any colony picker.
Singer Instruments (SI) offers a colony-picking platform, consisting of the ROTOR / The Stinger / PhenoBooth. The platform is versatile, customizable, and allows users to fully control every step of the picking process.
Here, we examine the picking accuracy of SI’s colony-picking platform by asking two questions:
Dr. Fred Cross at the Rockefeller University in New York designed a clever “checkerboard assay” for testing The Stinger’s ability to pick arrayed colonies. Using the assay, he had shown that the Stinger is able to pick with 100% accuracy when colonies are arranged in a 1536-density array.
How does 1536-density array translate to colony distances? The center-to-center distance between two colonies in a 1536 density array is 2mm. Given that the average diameter of a colony is around 1mm, we deduce that the edge distance for colonies in a 1536-density array is around 1mm (Figure 1). Of course, a colony bigger than 1mm would have less than 1mm edge distance between itself and its neighbours.
Putting together, we conclude that the minimum edge-to-edge distance required for SI platform to pick colonies accurately without contamination is 1mm.
Schematic diagram showing the distance relationships between colonies in a 1536-density array.
Unevenness of the agar or changes in agar height is a common factor causing a decrease in picking accuracy of a colony picker. This is because the distance between a colony and the camera, which directly affects colony position calculation, is a variable factor as a result of variation in agar height. For example, lower agar height causes colonies to be further away from the camera, making colonies appear to be smaller than what they actually are. On the other hand, higher agar height causes the colonies to be closer to the camera, making them look bigger (Figure 2).
The effect of agar height on colony size and position. A) Agar at “normal” height. B) Higher agar height: colonies appear bigger as they are closer to the camera. C) Lower agar height: colonies appear smaller as they are further away from camera.
The ROTOR is the fastest and most powerful colony manipulation robot in the world. It is essential for high-throughput screening and single colony picking.
In order to overcome this challenge, our colony picking platform was designed to account for changes in agar height across a plate. The Stinger uses the same floating head principal for surface detection as the ROTOR, and the PhenoBooth contains an agar depth which can be changed to reflect significant variation. To test its performance, we poured plates with different agar volumes into Singer PlusPlates. We then grew budding yeast cells on these plates. After 2 days of incubation, we picked 96 colonies from each plate. The accuracy was monitored by two methods. Firstly, we looked for pick marks of the colonies on the source plates. Secondly, we examined the growth of the corresponding colonies on target plates after incubation.
We found that our platform was able to pick with more than 98% accuracy when plates contain between 34 ml to 46 ml of agar, which corresponds to agar height between 4.7mm to 6.3mm. 40ml of agar fills a Singer PlusPlate half way up at 5.5mm height (Figure 3).
The Stinger picks with >98% accuracy at various agar heights.
In a separate test, we have found that unevenness in agar within a plate, probably caused by uneven bench surface during plate preparation, did not affect picking accuracy. We successfully picked 40 colonies at various positions on the plate in Figure 4 without a problem. This further confirms that SI’s platform is able to accommodate plates with different agar heights.
The picking accuracy is maintained through variations in agar height due to tools built into the Stinger and PhenoBooth. A) Side view of the plate showing agar height varies dramatically. B) Top view of the plate. C) Picture showing colonies selected for picking (pink colours).
The SI Colony-picking Platform, consisting of the ROTOR / The Stinger / PhenoBooth, can pick colonies with at least 98% accuracy when colonies are 1mm apart. Variations in agar height within the above tested range have minimal impact on picking accuracy (≥98%). Greater variations in agar height can be accounted for within the PhenoBooth Colony Picking software.
Contact us for more information about SI’s Colony-picking Platform.
See a randomly arrayed plate of colonies imaged with the PhenoBooth. PhenoSuite gathers the coordinates of selected colonies