Project Overview & Scope

2014 Experiment

The 2014 GxE Experiment consisted of a large hybrid trialing effort and a smaller inbred trialing effort. For our first year of field experiments, 25 principal investigators in 14 states and one Canadian province tested over 900 hybrids, deployed as groups of 250 hybrids with two replications at each location. With 23 participating locations, a total of 12,678 plots were grown. As part of the 250 hybrids, a set of 10 diverse hybrids was tested across all locations; other hybrids were selected based on reasonable maturity for specific locations. Except for the set of common hybrids, testers included LH198, LH195, LH185, CG102 and PB80. In addition, 31 inbred lines were replicated 2 to 12 times over 15 locations for a total of 1,598 plots. Inbred lines tested in 2014 include recombinant inbred lines (RILs), Expired Plant Variety Protection (exPVP) inbreds and lines from collaborator breeding programs.

---

2015 Experiment

For the second year of evaluations, the GxE experiment expanded to include Ohio, Kansas, and South Dakota, for a total of 17 states and one Canadian province overseen by 28 principal investigators. Data were collected from 28 hybrid experiments and 25 inbred experiments in 27 unique locations. For the hybrid experiment, effort was made to maintain continuity with 2014 while expanding the diversity of materials included in 2015 experiments and increasing overlap among locations. Most locations grew 500 hybrid plots, with a total of 13,650 plots. The design was modified from 2014 to increase the number of common hybrids to 45 hybrids in all locations plus 15 regional hybrid checks in Northern, Central, and Southern regions, each replicated twice per experiment for 120 entries. The remaining 380 hybrids were mostly unreplicated, except for the most northern and most southern locations where additional hybrids were included with two replications. Each hybrid was included in at least 8 locations and selected with consideration to adaptation, diversity, and replication of alleles. Except for the set of common hybrids, testers included PHB47, PHZ51, LH82 and LH195. For the inbred experiments, 32 inbred lines were replicated twice over 17 locations in 25 experiments for a total of 1,600 plots.

---

2016 Experiment

In 2016, the project continued to expand to Arkansas, Michigan, and South Carolina. In total, 40 hybrid experiments across 34 unique locations in 19 states and one Canadian province were overseen by 29 principal investigators. Significant growth in both the number of experiments, and the addition of several “mega locations” in Iowa, Wisconsin, and Missouri, resulted in 19,360 plots planted in 2016. New in 2016, the hybrid experiment was organized into five broad sub-experiments, summarized below:

1. Ex-PVP Factorial ("Design 2") Experiment: Planned ex-PVP diallels integrated with genomic prediction based on ex-PVP-derived progeny. Three sets of materials selected for appropriate maturity were planted in 28 locations. Led by Jode Edwards (USDA ARS), Elizabeth Lee (Guelph), and Martin Bohn (University of Illinois Urbana-Champaign).
2. Stiff Stalk GxE: Founder lines, public early releases, and ex-PVP lines, which represent a span of selection pressure, crossed by inbred line 3IIH6. The Stiff Stalk Set was planted in 17 locations. Led by Natalia de Leon and Shawn Kaeppler (University of Wisonsin-Madison).
3. Near-Isogenic Lines Experiment: Teosintes introgressions; B73/Mo17 near-isogenic lines (NILs). The NILs experiment was planted in 17 locations. Led by Sherry Flint-Garcia (USDA-ARS) and Candice Hirsch (University of Minnesota).
4. Germplasm Enhancement of Maize (GEM): A highly diverse set of materials containing un-adapted, exotic germplasm to enhance genetic diversity and trait performance, and to minimize risks to production. The GEM lines were crossed with inbred line LH195 and were planted in 17 locations.
5. Common Hybrid Checks: A set of 50 common check hybrids and five local checks were planted in all locations. Two smaller sets of regional common hybrids (15 early and 5 late) were planted in locations based on maturity. Hybrid checks were selected for diversity and continuity with 2014-2015 experiments.

---

2017 Experiment

In 2017, the project will progress further with the addition of three new principal investigators in Arizona, Colorado, and Mississippi. As a result, over 20,000 plots comprising 41 experiments are being grown by 30 investigators in 37 unique locations.The experiment in 2017 will replicate the five broad sub-experiments planted in 2016, with the addition of a sixth sub-experiment headed by Randy Wisser (University of Delawaree).

---

2018 Experiment

For planting season 2018, the experiment utilized genetic materials with a relatively narrow maturity window, across all locations; thus, resulting in reduced impact flowering time on results, presenting more uniform data, and simplifying plot management for the collaborators.

Experimental Design:

• Modified Randomized Complete Block Design (RCBD)

Description of Randomization:

1. Hybrids were grouped into four categories based on seed availability.
   1. Hybrids with 350 - 660 seeds = Group 4: tested in one rep at four randomly chosen environments.
   2. Hybrids with 660 - 1000 seeds = Group 8: tested in one rep at 8 randomly chosen environments.
   3. Hybrids with 100 - 1900 seeds = Group 24: tested in one rep at all environments.
2. Within a site, there are two replicates 'reps'. Any hybrids tested twice at a site appear once in each rep. So, for the subset of replicated hybrids, the reps are RCBDs. For unreplicated hybrids, hybrids where randomly assigned with stratification to one of the two reps. Stratification ensured nearly equal proportional representation of families in each rep, and about equal numbers of hybrids in each rep
3. Hybrids were grouped by families within reps (this grouping is labeled as 'block' in the design). Family order was randomized for each rep.
4. Check hybrids were randomly assigned to plots within reps, without regard to the 'block' structure.

Populations:

• PHW65 reference design population which includes double haploids derived from PHW65/PHN11, PHW65/Mo44, & PHW65/MoG

Testers:

• LH195 - Midwest to Southern Locations
• PHT69 - Northern Locations

2019 Experiment

In 2019, further continue testing with the same genetic material as in 2018. A new international collaborator joined the initiative, Georg-August-University from Göttingen, Germany. As a result, over 27,000 plots were evaluated across 31 unique locations. The experiment in 2019 used the same experimental design and testers adapted to each location.

2020 Experiment

For the 2020 experiments, 3 testers were decided to be used to have more closely adapted material to each location. For the first time select locations tested HIPS (High Intensity Phenotyping Site) plots with novel phenotyping methods and tools. The initiative also began to use UAVs (unmanned aerial vehicles) to measure specific traits in thousands of plots.

Experimental Design:

• Modified Randomized Complete Block Design (RCBD)

Description of Randomization:

1. Each trial is arranged in two replications (500 plots total). For the purposes of blocking in the field, the primary division is by replication (1 or 2) then by tester (PHP02, PHK76, PHZ51) if there are multiple testers. Most locations only included one tester.
2. Additional external blocks were created upon request for external Yellow Stripe replications to facilitate sampling or additional phenotyping. External experiments with 2 replications of 25 entries were also created upon request for Hybrid HIPS and Inbreed HIPS.
3. The objective of the experimental design was to balance the need for within-site replication against the overall goal of the GxE project to test as many different hybrids as possible at each trial site. If a location has hybrids with multiple testers, an equal amount of the plots were assigned to each tester as separate experiments.
4. Within an experiment there are two replications and each replication will have one plot of each of the core check hybrids (YS-Hybrids) based on seed availability.
5. Most hybrid trials are arranged in two-row plots, 20' long with 30-72" alleys between plots.

Population:

• Double Haploid lines were derived from the W10004 population.

Testers:

• PHZ51 -- Southern Location
• PHK76 -- Midwest and intermediate Locations
• PHP02 -- Northern Locations

---

2021 Experiment

2021 is the second year of evaluation of the genetic material present in 2020. The experiments used the same experimental design and testers. Additional smaller-scale experiments were created upon request to facilitate additional phenotyping for specific goals. Experiment = "YS" (External Yellow Stripe), "HIP_Hybrid", "HIP_Inbred", and "HIP_Hybrid_Low_Nitrogen".

---

2022 Experiment

For the 2022 experiments, the main group of hybrids evaluated in the main experiment 2022 (main experiment called "G2F_Main" in the phenotypic data) consists of doubled haploids that were testcrossed with ex-PVP inbred tester LH244. The doubled haploids were generated from F1 hybrids resulting from crosses between two groups of parents. One group of parents consisted of inbreds LH212Ht, PHJ89, PHK76, PHN46, and PHP02, which are members of the non-stiff stalk heterotic group. The second group of parents consisted of GEMN-0096, GEMN-0097, GEMN-0192, and GEMN-0225, which are populations released by the Germplasm Enhancement of Maize (GEM) project.

The experimental design was a modified randomized complete block design, with the objective of balancing the need for within-site replication against the overall project goal of testing as many different hybrids as possible at each trial location. A common set of 30 check hybrids was used to connect environments (combinations of location and year), and the check hybrids were fully replicated in each location (family called "CHECK" in the phenotypic data). The remaining hybrids were replicated 0, 1 or 2 times, depending on seed availability.

Additional smaller-scale experiments were conducted alongside the main experiment for additional phenotyping and/or deployment of novel phenotyping methods. These experiments included the external Yellow Stripe experiment (experiment called "YS" in the phenotypic data), which involved a set of hybrids know as "Yellow Stripe" hybrids that have been common across years. The High-Intensity Phenotyping Site (HIPS) experiment consisted of 22 hybrids (experiment called "HIP_Hybrid" in the phenotypic data) and 22 inbreds (experiment called "HIP_Inbred" in the phenotypic data).

---

2023 Experiment

In 2023, the evaluation continued with the same group of hybrids as in 2022. Same population, experimental design, and tester. Collaborators also requested additional smaller-scale experiments for phenotyping for specific goals and/or deployment of novel phenotyping methods.

---

2024 Experiment

For the 2024 experiments, the main group of hybrids evaluated in the main experiments (main experiment "G2F_Main_LH287", and "G2F_Main_PHP02") consists of doubled haploids that were testcrossed with ex-PVP inbred testers PHP02 and LH287. The doubled haploids were generated from F1 hybrids resulting from crosses between two groups of parents. One group of parents consisted of inbreds 3AZA1, FBLL, PHB47, and PHBW8, which are members of the stiff stalk heterotic group. The second group of parents consisted of GEMS-0061, GEMS-0113, GEMS-0219, and GEMS-0227, which are populations released by the Germplasm Enhancement of Maize (GEM) project.

Experimental Design: The objective of the experimental design for the main experiments is to balance the need for within-site replication against the overall goal of the GxE project to test as many different hybrids as possible at each Field-Location. Within an experiment there are two replications, and each replication will have one plot of each of the core check hybrids based on seed availability (Family = "CHECK"). A sample of at least 25 of the experimental hybrids is also replicated within each Field-Location (with a different sample replicated at each environment. Environment is a combination of Field-Location and year). The remaining plots are occupied by hybrids that occur in only a single replication within the environment. Finally, entries were assigned to incomplete blocks of 10 or 20 plots each within each replication. This represents a combination of features of incomplete block designs, augmented designs, and partially replicated designs. D-efficiency was optimized at each level of sampling in the design (selection of hybrids among environments, assignment to replications within environments, and assignment to incomplete blocks within replications) to maximize balance under the restrictions of variably limited seed availability for some hybrids.

---

List of core check hybrids for 2024-2025 GxE Project main experiments ("G2F_Main_LH287", and "G2F_Main_PHP02")

---

List of genotypes present in the "YS" (External Yellow Stripe), "HIP_Hybrid", and "HIP_Inbred" experiments.

---

Environmental Data Collected

Each experiment is furnished with a WatchDog 2700 Weather Station by Spectrum Technologies, Inc. The weather stations record air temperature, humidity, solar radiation, rainfall, wind speed and direction, soil temperature, and soil moisture every 30 minutes for the duration of the growing season. In addition, collaborators submit soil samples for basic nutrient and texture analysis to a central soil testing lab.

---

Phenotypic Traits

The following table summarizes phenotypic traits measured by each GxE experiment cooperator at each location.