top of page

Measuring Yield Together

Yield is the single most important trait measured in most contexts, as low yielding varieties result in more land being used and poor economic returns for farmers. We measure yield every year as part of the US GxE project (20+ locations per year), and we collaborate with breeders focused on yield globally for many species.


The last few decades have seen an increase in papers published claiming 10-68% increases in crop yields imparted through modifying one or a few genes. While genetic engineering, CRISPR-Cas9, and other genetic modifications have incredible potential to impact agriculture in light of climate change, small-scale trials claiming massive changes in crop productivity that do not deliver results may hinder progress in this space. Merritt Khaipho-Burch lead a Nature Commentary on "Scale-up trials to validate modified crops’ benefits", where we discuss common issues that arise within these studies and suggest ways that our community can move forward together to robustly test changes at scale.


Below, we detail many of the common issues that arise within small-scale studies so that researchers, reviewers, and journal editors can keep an eye out when evaluating this field of research. For a graphical version of these guidelines, find our downloadable flier below to decorate your lab or office space!


Read more within our Nature Comment article: 


Have questions? Please contact Merritt Khaipho-Burch (

Have ideas on building better public sector testing frameworks - reach out to Ed Buckler


Studies should use standard definitions of yield

  • Was plot yield measured as opposed to single plant yield or individual yield components (e.g., grain length, grain width)? Were these values scaled up with simplifying assumptions on planting density and area?


Trials should be replicated across plots, geographical locations, and years.

  • Were yield effects tested in the field as opposed to the greenhouse?

  • Was sufficient individual genotype and plot replication used? 

  • Were effects tested across different geographical locations (environments) and years?

  • Were field effects corrected for in yield estimates (BLUPs, BLUEs)?

  • Were yield estimates ‘cherry-picked’ in the main results to show the best-performing plots, individuals, or years?

  • Are yield effects stable across environments?


Varieties, planting densities, and other conditions should closely match those on farms.

  • Were yield effects tested, or backcrossed into elite or commercially competitive germplasm?

  • Were plots designed to avoid plant competition (i.e., border rows, four-row plots)?

  • Was a commercially relevant planting density used?

  • Were standard practices used that would be experienced on farms (i.e., practices around fertilizer application, tilling, irrigation, sowing, harvesting, etc.)?

  • Was grain or biomass moisture controlled to a commercial standard?


Appropriate controls should be used.

  • Was yield measured in a context relevant to what is commercially expected (i.e., hybrid versus inbred yield)?

  • Were yield effects tested in a non-transformed wild-type genotype?

  • Were yield effects tested in a null-construct/empty-vector genotype?


Researchers should prioritize genes that plant breeding might have missed. 

  • Is this gene or allele already present or fixed in commercial germplasm? If so, why hasn’t it been fixed by plant breeders? Is there a tradeoff?


Develop collaborations

  • Utilize publicly available yield testing opportunities to test changes at scale. These could include the Genomes to Fields Initiative, university-based experimental stations or breeding programs, and the Consultative Group on International Agricultural Research (One CGIAR). 

bottom of page