Fig. 4. Relative abundance of the TARA Oceans nifH genes in the context of reference collections and amplicons.

Violin plots summarizing the average mean coverage of nifH genes retrieved in this study, nifH reference databases^46,47 and nifH amplicon sequences from a large-scale survey³⁹ across 93 TARA Oceans metagenomes using a competitive read recruitment strategy. The 18 nifH genes retrieved in this study were separated into two groups (‘HBD genomes’ and ‘Orphan genes’ for which we only have a scaffold) and compared to a database of nifH gene sequences. For each gene sequence, the coverage values were corrected by excluding nucleotide positions with coverage in the 1st and 4th quartiles to minimize the effect of non-specific mapping.