​​Sequencing results
From the 110 bee samples submitted, 1 negative control, and 1 extraction liquid sample for PacBio long amplicon sequencing, 532,208 raw reads and 326,739 high-quality sequences were retained after processing. From these sequences, 810 unique amplicon sequence variants (ASVs) were inferred, 86 of which were classified as chloroplast or mitochondria and removed. The samples containing fewer than 500 total sequences (17, all crop and extraction or negative controls) were removed. Overall, the pipeline resulted in a final dataset containing 724 ASVs and 93 samples (55 gut and 38 crop) with a median of 2,472 sequences per sample. Sampling curves (Supplementary Figure S2) were saturating, indicating sufficient sampling depth.