• Reduce text

    Reduce text
  • Restore text size

    Restore text size
  • Increase the text

    Increase the text
  • Print


Phenomic selection: an opportunity for the future?

Genomic selection has enabled the improvement of breeding schemes in numerous species where the genotyping matrix serves as a foundation to predict the performance of new individuals. Less costly and equally as efficient, phenomic selection – based on near infrared spectroscopy – will offer an inexpensive alternative for selection in many species.

La sélection phénomique : une voie d’avenir en sélection ?. © INRA
Updated on 05/22/2019
Published on 04/18/2019

The principle of phenomic selection

Phenomic selection is a new technique for varietal breeding based on indirect predictions. The principle is similar to that of genomic selection but the varieties are characterised using near infrared spectroscopy instead of genotyping. Recent research has indeed demonstrated that other sources of molecular variations (such as those resulting from transcription or metabolites) could be of value to accurately predict complex traits. Obtained using a high-throughput method (near infrared spectroscopy) these data capture genetic information and therefore enable estimates of the similarities between individuals so that it is possible to produce predictions of the genetic profile of new varieties. The technique is based on measuring reflectance at different wavelengths on the tissue under study (composition in metabolites, proteins, etc.). This reflectance is mainly linked to vibrations in the chemical bonds between the molecules making up the tissue.

A proven approach in two species of economic interest: wheat and poplar

The principle consists in calibrating a prediction formula on phenotyped reference material and then predicting the performance of new candidates. Applied to a variety of tissues (seeds, leaves, wood) in two species – soft wheat (Triticum aestivum L.) and poplar (Populus nigra L.) – the method has demonstrated a degree of reliability equal to that of molecular markers in terms of predicting traits for development, disease tolerance and yield in new varieties. The prediction formula was standardised on a panel of 228 elite European varieties, most of which had been on the market since 2000. For poplar, research was based on an association population of 1160 cloned genotypes representative of the natural distribution of this species in Western Europe. For both species, phenomic selection demonstrated good predictive capacities, even when the measurement environment differed radically from that where the spectroscopy data were collected.

The advantages of phenomic selection and perspectives for breeders

High-throughput near infrared spectroscopy is an inexpensive tool. While the costs of genomic selection are around €35 and €50 per individual for wheat and poplar, respectively, those of phenomic selection are closer to €3 per analysis for wheat and €2.5 for poplar. This method will also serve as an alternative to genomic selection in numerous orphan species where genotyping tools are not available at a reasonable cost.
Because it is not destructive and the samples do not require any particular treatment prior to analysis (no DNA extraction), phenomic selection could be applied to any plant or animal species. It is now possible to obtain the spectral signature of tissues directly in the field, thanks to portable systems or high-throughput autonomous vectors such as phenomobiles that generate hyper-spectral images. It is also possible to determine the spectra directly on seeds, so that predictions and selection regarding the best candidates could be performed before the initiation of trials. The technological advantages anticipated in terms of automating the collection of high-throughput reflectance data at an industrial scale and a concomitant fall in costs will further enhance the advantages of phenomic selection and offer new perspectives for the breeding community.

Partnership and Innovation Manager:
Thomas Goujon (+334 72 72 89 83)

Find out more

The teams involved

INRA Val de Loire

  • Experimental Unit for Forest Genetics and Biomass - Orléans (GBFor)
  • Génobois Platform, Orléans: platform for phenotyping the physicochemical properties of wood and the hydraulic properties of trees
  • Joint Research Unit for Forest Tree Breeding, Genetics and Physiology (BioForA)

INRA Versailles – Grignon

  • US EPGV: Study of the Polymorphism of Plant Genomes (CEA/Genomics Institute/National Genotyping Centre Evry)

INRA Clermont-Ferrand

  • Phéno3C: Platform for Genotyping in the Field


  • Experimental Unit on Arable Crops, Innovation and the Environment – Picardie (UE GCIE)


  • Genotoul Toulouse: Network of technology platforms

This research work benefited from financial support from France Agrimer and under the following projects:

The Phéno3C Platform was funded by the French National Research Agency (ANR) in the context of the Phenome project under the French government’s Investments for the Future programme (ANR-11-INBS-12), and by the European Regional Development Fund (AV0011535).