Set | City | Country |
|---|
New Zealand | U.S.A. | Nigeria | Portugal | Chile | Japan | Colombia |
|---|
Training Main | Auckland (AKL) | 15 | | | | | | |
Hamilton (HAM) | 16 | | | | | | |
New York (NYC) | | 26 + 0 = 26 | | | | | |
Offa (OFA) | | | 20 | | | | |
Porto (PXO) | | | | 60 | | | |
Sacramento (SAC) | | 16 + 18 = 34 | | | | | |
Santiago (SCL) | | | | | 20 | | |
Tokyo (TKO) | | | | | | 20 | |
Testing Mystery-1 | Various (C1) | | 10 (NCY) | 5 | 10 | 5 | | |
Training Mystery-2 | Ilorin (C2) | | | 12 | | | | |
Lisbon (C3) | | | | 12 | | | |
Boston (C4) | | 12 | | | | | |
Bogota | No samples in the training set |
Testing Mystery-3 | Various (C5) | | 3 (Boston) | 4 | 4 | | | 5 (Bogota) |
- Table also shows the mystery sets and how the city and sets were internally coded in this work. The column corresponding to US, shows that samples from New York City and Sacramento included additional samples from the pilot analysis but those samples yielded OTUs in this present setting only in Sacramento. Light gray rows are multi-city groups where city of provenance was predicted (testing sets) for all the samples according with the corresponding training model (main or mystery-2). All samples in training sets had a counterpart in the testing sets with the exception of the city of Bogota, which has 5 samples in the testing set (mystery-3) but has no samples in the training set (mystery-2)
