Tool predicts plant reactions

A new computing tool could help scientists predict how plants will react to different environmental conditions in order to create better crops, such as tastier and longer-lasting tomatoes.

The tool will form part of the £1.7m Syngenta University Centre at Imperial College London, where researchers from Imperial and Syngenta will work together to improve agricultural products.

Scientists around the world have aspired to develop strains of crops such as drought-resistant wheat and pesticides that are more environmentally friendly. In order to do this, however, they need to predict how the genes inside plants will react when they are subjected to different chemicals or environmental conditions.

Prof Stephen Muggleton, director of the centre from the department of computing at Imperial College London, said: ‘We believe our computing tool will revolutionise agricultural research by making the process much faster than is currently possible using conventional techniques. We hope that our new technology will ultimately help farmers to produce hardier, longer-lasting and more nutritious crops.’

The researchers have developed a prototype of the tool, which they are currently testing. It is claimed that it can analyse in a matter of minutes, instead of months, which genes are responsible for different processes inside a plant and how different genes work together.

The tool uses a type of computer programming that relies on machine learning, a set of sophisticated algorithms that allows a computer to learn based on the data that it is analysing. According to the researchers, the tool will recognise complex patterns in that data to find pieces of information about plant biology that might have previously taken months or even years to find.

The machine-learning ability of the tool means that researchers can develop an understanding of different plants, even when they are lacking information about some aspects of their inner workings.

The researchers claim that, previously, mathematical modelling of a plant's behaviour was time consuming and difficult because, without all the information about a plant, the models were imprecise.

For the first project using the tool, scientists will look at how different genes affect the way that a tomato's flesh hardens and tastes and how the fruit's skin changes colour from green to red.

The researchers hope that this will enable them to develop tomato strains that are tastier and that redden earlier and soften later so that they can be transported more easily to market. It is claimed that these qualities could be especially useful in developing countries, where factors such as poor transport can quickly spoil fruit and vegetables.

The team is also proposing another project to test the safety of pesticides that Syngenta is developing, before they reach the manufacturing stage. The tool will allow the researchers to construct models that reveal, for example, whether a proposed pesticide might affect metabolites, which are responsible for processing energy inside a plant.

All software developed by researchers at Imperial College is intended to be made publicly available over the next four years.