On one level, RamSelect is a tool for choosing rams based on genetic traits. On another, it is making an important point about the value of the user interface - the window through which people access otherwise bafflingly complex technologies.
The Netscape browser put a face on a little-understood thing called the internet. The iPhone encased a lot of geeky technologies, most of them invented for other purposes, and allowed the user to manipulate them all with a finger. With simple user interfaces, Uber and AirBnB have been able to build networks that link basic needs (transport and shelter) with underutilised assets (private cars and homes).
RamSelect probably won’t make it into this album of technological Greatest Hits, but the principle is the same: it has put a simple user interface between sheep producers and the dizzyingly complex science of genetic selection.
RamSelect has a twofold purpose.
One is to allow sheep producers to data-mine the immense databases of Sheep Genetics for statistical information that helps drive better breeding decisions. The other purpose is to make that process as simple and useful as possible.
Simplicity has not been a feature of Australian Sheep Breeding Values (ASBVs), the currency of sheep genetics.
For some ASBVs, a high number is good; for others, high numbers are bad. Every trait interacts with others in ways that the numbers don’t represent.
To get around this complexity, geneticists compiled indexes: ideal combinations of traits intended to boost productivity while minimising antagonism between traits.
RamSelect puts a new face on all these numbers and indexes, and makes the machinery of genetic selection invisible.
Instead, the app presents producers with sliders they can move to emphasise or de-emphasise certain traits. When the sliders are positioned, the app provides a list of rams available for sale across participating studs, ranked according to how well they fit the criteria indicated by the sliders.
The development of the RamSelect app involved months of debate between Sheep CRC geneticists and app developers, Pivotal Labs in San Francisco.
Sheep CRC chief executive, Professor James Rowe, said the debate illuminated the differences between what scientists think is important, and what interface designers think is important.
An example: the scientists wanted trait correlations reflected in the sliders. If, for instance, the wool quality slider was moved one way, the scientists felt the other sliders should be automatically re-indexed according to how each trait interacted with wool quality genes.
Pivotal successfully fought against this, Dr Rowe said. If the user moves one piece of the interface, and other bits then move, “that suggests that there is some funny business going on in the background. Pivotal wanted absolutely straightforward interaction with this app”.
Both sides won. Genetic interactions are still displayed in RamSelect, only in a pie graph that graphically portrays how much emphasis the user is placing on each trait.
RamSelect v2 is in the works, and Dr Rowe is animated about how it should enhance v1’s ease of use, while adding depth and value.
The interface is being refined “to give clear sense of where you are in the full spectrum of rams available”.
At the same time, value is being deepened by allowing producers to log their own flock genetics as a baseline for decisions. They will be able to enter inventories of rams they have bought using RamSelect (and eventually, genomic information from their own ewe flocks), so that each trait slider in the app shows the user’s own genetic position in relation to all the rams in the database.
RamSelect Version 2 is due out by mid-2016.