statistical design

Taking the kinks out of curves

In a recent paper, researchers have developed a methodology suitable for analyzing the growth curves of a large number of plants from multiple families. The corrected curves accurately account for the spatial and temporal variations among plants that are inherent to high-throughput experiments.


An example of curve registration.  a The salinity sensitivity (SS) curves of the 16 functions from an arbitrary family, b SS curves after the curve registration, and c the corresponding time-warping functions. The salinity sensitivity on the y-axis of a and b refers to the derivative of the relative decrease in plant biomass


Advanced high-throughput technologies and equipment allow the collection of large and reliable data sets related to plant growth. These data sets allow us to explore salt tolerance in plants with sophisticated statistical tools.

As agricultural soils become more saline, analysis of salinity tolerance in plants is necessary for our understanding of plant growth and crop productivity under saline conditions. Generally, high salinity has a negative effect on plant growth, causing decreases in productivity.  The response of plants to soil salinity is dynamic, therefore requiring the analysis of growth over time to identify lines with beneficial traits.

In this paper the researchers, led by KAUST and including Dr Bettina Berger and Dr Chris Brien from the Australian Plant Phenomics Facility (APPF), use a functional data analysis approach to study the effects of salinity on growth patterns of barley grown in the high-throughput phenotyping platform at the APPF. The method presented is suitable to reduce the noise in large-scale data sets and thereby increases the precision with which salinity tolerance can be measured.

Read the full paper, “Growth curve registration for evaluating salinity tolerance in barley” (DOI: 10.1186/s13007-017-0165-7) here.

Find out how the Australian Plant Phenomics Facility can support your plant science research here.


High-throughput phenotyping in the Smarthouse™ at the Adelaide node of the APPF

NW 24-7-15 (2)

Barley plants growing in the Smarthouse™



A better way to tackle environmental variation in your greenhouse research

Statistics prove the smart way to deal with variation in your controlled environment greenhouse.

Plant phenomics allows the measurement of plant growth with unprecedented precision. As a result, the question of how to account for the influence of environmental variation across the greenhouse has gained attention.

Controlled environment greenhouses offer plant scientists the ability to better understand the genetic elements of specific plant traits by reducing the environmental variances in the interaction between genetics and environment.

But controlled environments aren’t as controlled as they seem – variation does exist. For example, some days are cloudy, some are not. The sun, as it crosses the sky, casts shadows differently on plants, depending on their position within the greenhouse. In fact, a recent study by colleagues at INRA in Montpellier showed significant light gradients within a greenhouse and provided sophisticated tools for understanding how much light each plant receives.

One practice for dealing with variation has been to rearrange the position of the plants around the greenhouse during the experiment, however, there is a better way.


Rice plants growing in The Plant Accelerator® at the Australian Plant Phenomics Facility’s Adelaide node

The automated high-throughput phenotyping greenhouses at The Plant Accelerator® are controlled environment facilities which use sensor networks to identify and quantify environmental gradients (light, temperature, humidity) in the greenhouses. To further tackle environmental variation, Chris Brien, Senior Statistician at The Plant Accelerator®, led a study that showed good statistical design and analysis was key to accounting for the impact of environmental gradients on plant growth. It was argued that rearranging the plants during the experiment makes it impossible to adjust for the effect of gradients and should be avoided.

The study involved a two-phase wheat experiment involving four tactics in a conventional greenhouse and a controlled environment greenhouse at The Plant Accelerator® to investigate these issues by measuring the effect of the variation on plant growth.

To learn more about Chris’s study read the full paper here.

To discuss the benefits of good statistical design contact Chris Brien.

To access The Plant Accelerator® for your research:  The Plant Accelerator® at the Australian Plant Phenomics Facility (APPF) is available to all publicly or commercially funded researchers. We have a full team of specialists including statisticians, horticulturalists and plant scientists who can provide expert advice to you when preparing your research plans.