Fig. 2

Hierarchical clustering of samples using taxonomic profiles at the genus level (R1 dataset). The taxonomic profiles clustered based on sampling location. The color scale reflects log-normalized proportional values (e.g. −1 ~ 10 %, −2 ~ 1 %, −3 ~ 0.1 %). JPL-SAF samples clustering together and showing several unique low abundance members were not found in the JPL-103 Debris and ISS samples (e.g., Mollicutes, Nitrospira, and members of Chloroflexi). Rows and columns are clustered independently using the furthest neighbor algorithm with a Euclidean distance metric. o order, f family, g genus