The goal of the research project is to increase the consistency and efficiency rates for the microinjection of embryonic stem cells into blastocysts through automation and development of an intelligent control strategy.
Interaction with the Animal Model Core Facility at UNC-Chapel Hill showed that cell micro-injection systems needed transformation, from a manual, time-consuming, and tedious, micro-injection process to an automated process. As a first step towards achieving full automation we have assembled a system that is semi-automated. The major feature of our semi-automated system is an interactive user interface, one that treats the cell microinjection process like a computer game. From the strategies adopted by many players we will devise a controller based on the experimental results and machine learning techniques.
In Figures 1 and 2 the system for positioning the injection needle under user control is displayed, using a joystick. Figure 3 shows how video processing software was developed to find a blastocyst, and how to manipulate and record the position and orientation of the blastocyst using the injection needle. The goal is to have the blastocyst located at the holding pipette with the inner cell mass positioned at 180 degrees from the stem cell injection needle.
Tests showed that an expert could improve their rate of successfully injecting cells, from 40%-70%, to 80%. More, a novice, with no knowledge of cell in vitro fertilization could also get 80% success with two days training on this system.