Most in silico evolutionary studies commonly assumed that core genes are essential for cellular function, while accessory genes are dispensable, particularly in nutrient-rich environments. However, this assumption is seldom tested genetically within the pangenome context. Here we conducted a robust pangenomic Tn-seq analysis of fitness genes in a nutrient-rich medium for Sinorhizobium strains with a canonical open pangenome. The Hidden Markov Model-based method is the most robust to variations among mutant libraries per strain and non-sensitive to data size. Fitness genes, categorized as essential (ES), advantage (GA) and disadvantage (GD) for growth, are enriched in core genes, while nonessential genes (NE) are over-represented in accessory genes. Connectivity degrees in predicted protein-protein interaction networks decline in the order of ES, GA, GD and NE, and those in the co-fitness network decreases from ES to GA/GD to NE. Notably, 1599 out of 3284 core genes display differential essentiality across test strains. Within pangenome core, both shared quasi-essential (ES and GA) and strain-dependent fitness genes are enriched in similar functional categories. Our analysis demonstrates a considerable fuzzy essential zone determined by co-fitness connectivity degrees in Sinorhizobium pangenome, and highlights the power of co-fitness network in understanding genetic basis of ever-increasing prokaryotic pangenome data.