Skip to main content

Table 2 Comparison of misclassified isolates with Random Forest to traditional laboratory testing

From: Genome-wide association studies of Shigella spp. and Enteroinvasive Escherichia coli isolates demonstrate an absence of genetic markers for prediction of disease severity

IsolatePhenotypeaRandom Forest (RF)aVotesbLocation in SNP treeSerotype Shigella/E. coli (agglutination)Properties against RF classification
IBESS811ES0.99Within
S. sonnei
S. sonnei phase 1/ O-negativeMotility
IBESS97ES0.80Within
S. flexneri
S. flexneri, inconclusive/ O135Inconclusive Shigella serotype
IBESS1163ES0.76Within
S. flexneri
S. flexneri, inconclusive/ O135Inconclusive Shigella serotype
IBESS911ES0.68Within
S. flexneri
S. flexneri, inconclusive/ O135Inconclusive Shigella serotype
IBESS996SE0.53Within EIEC / S. flexneriS. flexneri 3a/ O135None, hybrid isolated
IBESS988SE0.56Within EIEC / S. flexneriS. flexneri 3b/ O135None, hybrid isolated
IBESS419SE0.57Within
S. flexneri
Provisional/O-negativeNone, hybrid isolate, provisional Shigellad
IBESS232SE0.60Within
S. flexneri
Provisional/O-negativeNone, hybrid isolate, provisional Shigellad
IBESS470SE0.82Within EIECProvisional/O-negativeNone, hybrid isolate, provisional Shigellad
IBESS810SE0.89Within EIECAuto agglutinablecNone, hybrid isolate, provisional Shigellad
  1. RF Random Forest. aE Escherchia, S Shigella. bfraction of votes for classification in Random Forest. cIn-silico serotype, using E. coli serotypeFinder 2.0 of the Center for Genomic Epidemiology [23]: provisional/O-negative. d Hybrid isolates Isolates that possess characteristics of both Shigella spp. and E. coli.