Unlike the widely used butterfly-like all-reduce algorithm that incurs network contention in SMP/multi-core clusters, the proposed algorithm can achieve contention-free communication in almost all contemporary clusters, including SMP/multi-core clusters and Ethernet switched clusters with multiple switches.
The genes that serve as 'signature' genes for the subtypes are significantly differentially expressed between pairs of core clusters representing the subtypes.
Similar results were obtained when confining the analyses to just the core clusters, i.e., the clusters with the largest numbers of members, where there was 98% (2,122 of 2,266) agreement between the two methods.
To estimate the fraction of the group core genomes affected by recombination within the "Firm-5″ and "Bifido" groups, the genes of the group core clusters were aligned individually at the protein level with mafft.
The core clusters are collected together in the set (Line 15).
We did not take into account core clusters (which are the same for all strains) and unique clusters.
Even in this reduced set of 186 strains, 3 other strains (CZ383, P533H and R3262; all carriage strains isolated from the nasopharynx), each in a different lineage, are each missing over 50 core clusters based on the 24 WGS strains.
When we applied iNMF to 10.000 additional randomly selected probe set groups, the core clusters did not change significantly showing that the discovered subtypes do not depend on the randomly selected probe set groups.
The three organisms share 571 core clusters, somewhat more than the conserved crenarchaeal core of 336 determined by Makarova et al. [ 22].
The core clusters D in are then sorted decreasingly according to the number of collections among in which cluster D is contained (Lines 18-19), where ties are broken arbitrarily.
