Durum wheat (Triticum turgidum L. ssp. durum) is one of the major staple crops in the world, being the main source for the production of pasta, semolina and couscous. The ever-growing food demand of an increasing world population has brought in the need of an improved grain yield. Up to now the majority of the yield improvement has been obtained through increasing the number of spikes per area and the grain number per spike. Spike traits like fertile spikelet number, spike length, sterile spikelet number and floret fertility can determine the number of grains per spike and other quality-related traits, therefore they might as well affect quality and yield potential.

The Global Durum Panel (GDP) is a wheat collection composed mainly of durum modern cultivars and durum landraces, for a total of nearly 800 accessions. The Tetraploid Global Collection (TGC) consists of approximately 1800 genotypes, comprehensively sampling wild emmer wheat, domesticated emmer wheat, durum landraces, and other durum subspecies. Both panels were genotyped using the wheat high-density Illumina iSelect 90K SNP assay to provide a common genotype framework.

In this ongoing study, GDP and TGC were evaluated in multiple field trials in different environments over two seasons. The field trials were located in Cadriano, Fiorenzuolad’Arda (Northern Italy) and Grosseto (Central Italy) in 2019-2020-2021. The collections were grown in plots of 1 square meter under an unreplicated modified augmented design divided in rows, columns and blocks. Checks were randomly distributed in each block in the field layout.

For each accession, six spikes were characterized for several yield-related traits (i.e. average spike length, sterile spikelet number, fertile spikelet number, number of fertile florets per central spikelet, number of sterile florets per central spikelet). Some traits showed a remarkable variability in each trial such as fertile spikelet number (from 12 to 41) and number of fertile florets per central spikelet (from 1 to 12). This variability has been highlighted especially in the TGC, thus indicating a higher range of diversity in the ancestral wheat.

A preliminary GWAS was conducted for the GDP with R package GAPIT3 using different models (GLM, MLM, MLMM, FarmCPU and BLINK) and including the kinship matrix K. Especially for the number of fertile florets per central spikelet, the tag SNP “BobWhite_c46780_84” was detected with three different models (GLM, FarmCPU and Blink) on chromosome 2A in the 130.385.623 bp position as associated to increased number of fertile florets in spikelets. This could be a putative QTL for Grain Number Increase in the short arm of chromosome 2A with the favorable allele introduced by CIMMYT breeding program.

A more complete GWAS will be performed for both GDP and TGC including population structure in order to detect novel loci involved in yield and its components.