Fig. 3: Comparison of contig and gene prediction metrics between co-assembly and individual assembly.

Differences in a total contig number, b total contig length, c total gene number, and d total gene length between the two assembly approaches were evaluated using paired one-sided Wilcoxon signed-rank tests. Boxplots display median values as center lines, with lower and upper hinges representing the 25th and 75th percentiles. To ensure comprehensive data representation, whiskers extend to the full range of observed values, including outliers. e–g depict the cumulative distribution of gene lengths for complete, partial, and incomplete genes, respectively, across the two assembly approaches. For visual clarity, genes in each sample were divided into 50 bins, each bin containing an equal number of genes sorted in descending order of length. Within each bin, total gene length and the number of genes were calculated.