Fig. 2

Country and continent drive microbiome structure. a PCoA of Bray-Curtis distances between all samples. b PERMANOVA based on Bray-Curtis distances between CRC and non-CRC control samples and UK volunteer samples. For the CRC/non-CRC control samples: ‘country of origin’ = Argentina, Chile, India or Vietnam; ‘disease status’ = CRC or non-CRC control. For the UK volunteer samples: ‘country of storage’ = Argentina, Chile, India, Vietnam or UK (only samples stored in the UK or short-term outside the UK were used in this analysis, as these samples underwent DNA extraction at the same time); ‘duration of storage’ = short-term or long-term storage outside the UK (samples which remained in the UK were excluded from this analysis). **p ≤ 0.01. *p ≤ 0.05. c Distribution of Bray-Curtis distances for CRC and non-CRC control samples: ‘disease-status group’ = CRC or non-CRC control. d Shannon diversity indices for CRC and non-CRC control samples. At the taxonomic resolution achievable from the data, Asian individuals generally had lower diversity than South American individuals. e Network representation of all samples based on Bray-Curtis distances. Each UK volunteer is represented by a single point, reflecting the average relative abundance across all samples derived from that individual. Overall dissimilarities are driven by a weak segregation between Asian and South American microbiomes