Comparison of phenotypic selection of inbred lines, genomic selection of inbred lines, and evolutionary populations for field pea breeding in three Mediterranean regions

Paolo Annicchiarico^1*Meriem Laouar²Imane Thami-Alami³Margherita Crosta¹Nelson Nazzicari¹Luciano Pecetti¹Luigi Russi⁴

• ¹Council for Agricultural Research and Agricultural Economy Analysis | CREA, Rome, Italy
• ²Ecole Nationale Supérieure d'Agronomie, Algiers, Algiers, Algeria
• ³National Institute for Agricultural Research (Morocco), Rabat, Morocco
• ⁴Dipartimento di Scienze Agrarie, Alimentari e Ambientali, Università degli Studi di Perugia, Perugia, Umbria, Italy

Pea breeding may rely on phenotypic selection (PS) of single-seed descent (SSD) or bulk-derived lines, line genomic selection (GS), and selection of evolutionary populations (EPs). This study aimed to compare all these approaches in region-specific grain yield selection for Central Italy, coastal Algeria, and inland Morocco, using material from three connected crosses of elite, geographically diverse parent cultivars. Bulk breeding and initial EP development were based on three-year mass selection for plant yield under managed severe drought for Algeria and Morocco, and four-year field-based natural and mass selection under autumn sowing in Northern Italy for Central Italy. One EP per region was eventually developed by additional region-specific three-year field-based natural selection on pooled three-cross material. Region-specific line selections were performed on each cross using three experiments per region for PS, and GS models developed in previous studies. An additional GS was performed for a putative Stressful Italy region by combining predictions for Italy and Morocco. In a comparison of top-performing bulk-derived lines vs. SSD-derived lines, the former out-yielded the latter by at least 10% in four environments featuring the same predominant stress (drought or low winter temperatures) that had acted on bulked progenies. All selections were evaluated in one location per target region for two cropping years, in a managed drought environment, and in intercropping with barley in Algeria and Morocco. Results indicated that: (a) EPs did not differ from GS-or PS-derived lines for mean yield in their target regions while showing greater yield stability, better response to a climatically unfavourable year, and broader adaptability to intercropping or other non-target environments; (b) EPs were out-yielded by the top-yielding line in each region; (c) GS-and PS-derived lines did not differ in mean yield or yield stability, but a superiority of GS over PS emerged for Algeria and tended to emerge for Morocco when comparing the top-yielding lines; (d) GS-derived lines for Stressful Italy displayed comparable mean yield and higher yield stability than other region-specific GS-derived lines. Our results suggested different opportunities for adopting and integrating biotechnology-and agroecology-based approaches depending on the characteristics and objectives of the breeding program.