Robotic arm with a small foot-print

Master's degree/engineering course project

(The earliest starting date June 1st 2018. For more info contact Alyona)

 Fig. 1. Six days of time-lapse tracking of the root growth. Arabidopsis thaliana seedlings
 grown on a Petri dish in a plant growth chamber. 

The research in our group is focused on molecular mechanisms underpinning development of plants. It is similar to engineering when there is nobody around to explain you what all the parts are for. So, on daily basis, we break plant genes to see what they were needed for. To track plant growth properly we need a good camera system that will image plants for days or even weeks (exactly like Fig 1, but less blurry). We grow our plants in small Petri dishes within a growth cabinet and thus need the imaging platform to have small foot print.

Preliminary results:

We put together the minimal version of the imaging stage (see Fig. 2). Despite the questionable quality of the design, it was such a significant improvement of our experiments that we decided to invest into constructing a proper system. Our system is built around a Raspberry Pi computer, connected to a camera capable of producing images both under daylight conditions and at night using near-infrared illumination.

The growth cabinet

The imaging stage

Fig. 2. The current layout of the imaging stage.  The imaging takes place within the plant growth cabinet, which maintains optimal temperature and light intensity for the plants. Plants within a Petri dish are mounted on a Sugru holder and imaged by the IR camera taped to the piece of cardboard. The camera is plugged into a raspberry pi computer that saves images on a server connected via WiFi.

Our group has just installed a Prusa i3 MK3 3D printer!

In collaboration with JJ we will soon print our first prototype for time-lapse imaging  of 4 plates.

Fig. 3. Update (August 2018).  No more double-sided tape on cardboard! We are switching to 3D printing.

Project goals and requirements: 

1. Make the imaging stage less prone to falling apart. Additionally, the stage should be constructed in such a way that it provides more consistent images with regards to lighting.

2. Develop a small footprint robot that can move plates from the growth position to the imaging stage and back. This will enable parallel imaging of multiple plates, while not compromising the amount of light available to the plants.

3. The robot needs to be able to communicate with the Raspberry Pi, either via ethernet, WiFi, bluetooth or serial/GPIO (or you may have other suggestions). 

Why should you join us:

Fig 3. The team.

1. You will work in a fantastic team (see Fig. 3). We love what we do, and call it "work" just because we get paid for it.

2.  You will co-author an open access publication describing the hardware and software of the system, helping other scientists to build on what we develop. All blueprints and software packages will be made available on-line under a permissible open source license.

3. We will introduce you to engineering on the DNA level.

Seed Plating Robot

15 HEC Engineering project

Background

Our group works with a plant model organism called Arabidopsis thaliana. For some of our projects we image seeds of Arabidopsis placed on the top of agar-containing medium in Petri dishes. Putting seeds on plates is a boring and tedious task and we want to develop a robot that could do it for us. The robot should be able to plate seeds on two types of Petri dishes (round, ⌀ 9 cm dishes containing 25 ml of the agar-containing medium, square 12x12 cm Petri dishes containing 50 ml of the agar-containing medium). Ultimately, we would like to publish the results of this project.




Figure 1. Suggested design, any modifications/optimizations and alternative designs are most welcome.
A. Suggested design. 1. Indentation in the stage to accommodate 9 cm round Petri dishes. 2. Indentation in the stage to accommodate 12x12 cm square Petri dishes. 3. Holders for 1.5 and 2 ml eppendorf tubes containing seed stocks. 4. Holders for sterile disposable tips. 5. Dispensing head that can transfer a single seed from a stock on a designated place onto a plate. 6-8. Movement in three dimensions should be designed to minimize the robots footprint and possible contamination of the medium on the plates. B. Suggested grid layout for coordinating seed transfer on plates.  Please note, that it is important to introduce asymmetry (one extra seed? A dot-like scratch on the medium?) to denote the left top corner. C. Seeds transferred on the plates will be later imaged using our PetriPi robot, images will be processed using automated image analysis. Seeds will germinate and roots will grow gravitropically, thus there should be at least 2 cm distance from a seed row to the bottom rim of a plate.




Hardware to be developed in this project

• Smallest possible footprint robot that can transfer seeds from the stocks in 1.5 or 2 mL eppendorf tubes onto plates.
• Seed plating must be done under sterile conditions (in a Fortuna Clean Bench), thus robot parts have to be autoclavable (120 ᵒ C), resistant to 70% Ethanol and not shedding particles on open plates
• Seeds must be put on the surface of the medium. For automated analysis of images it is important that there are no damages on the agar around the plated seed. Seeds can be dispensed while being dry (please note, that dry seeds are more difficult to handle due to static electricity) or they can be suspended in sterile water. If needed, the robot can be connected to vacuum supply, alternatively, we can also provide an old Gilson pipette that can be used to pipette a fixed volume and discard tips.
• Seeds should be plated in rows, with at least 3 mm distance between the seeds and at least 2 cm vertical distance between the rows. There should be asymmetry in the plating pattern that will allow to identify left top corner. Optimally it would be great to develop a robot that will confirm successful plating of a seed before proceeding to the next.
• Stocks should not be contaminated with seeds that don’t belong to them, thus it is desirable to use disposable tips for the seed dispenser. These can be custom-designed, alternatively there are commercially available filter tips (we can provide samples).
• The minimal number of seed to be plated from each stock = 4. Thus it can be used as a single input unit, when user defines coordinates for plating.
• The stage should enable fixed positioning of two types of plates.
• Plating time must not exceed 40 minutes, as it will lead to drying the medium.

Software to be developed in this project

• The area of Petri dishes should be split into a grid. Single input unit for plating (e.g position A1 on the Fig1B.) should correspond to positioning 4 seeds from one stock.
• User should be able to define seeds from what stock should be plated into which square
• The robot can be controlled via any method that is sufficiently user friendly, e.g. Ethernet connection to user’s compute, SD card, USB stick
• Software should predict the time required for plating. An optional feature would be to also notify the user that plating is complete.
• The software must be open-source

Facility

• Prusa i3 MK3 3D printer is readily available in our group
• All running costs will be covered by our group. Please provide us with invoices!

Grading

• All hardware and software developed in this project should be available to our group
• A comprehensive report should be provided by the end of the project.
• Top grade: the robot transfers more than 90% of seeds from a single stock on two types of plates from at least two stocks
• Medium grade: the robot transfers at least 90% of seeds on two types of plates from a single stocks
• Low grade: the robot transfers some seeds from a single stock

Main supervisor:

• Pernilla Elander, PhD student, Dept. of Mol Sci, SLU e-mail: pernilla.elander (at) slu.se

Co-supervisors:

• Jonas Ohlsson, PhD student, Dept. of Mol Sci, SLU e-mail: jonas.ohlsson (at) slu.se
• Alyona Minina, Ass. Prof.,  Dept. of Mol Sci, SLU e-mail: alena.minina (at) slu.se