Table 3 Selected markers using stepwise2 procedure in phylum and genus level. The mean train and validation AUCs are recorded.

From: Machine learning models for pancreatic cancer diagnosis based on microbiome markers from serum extracellular vesicles

(A) Phylum

# markers

marker

AUCtr

AUCval

0

class ~ sex + age

0.599

0.545

1

+ Verrucomicrobia

0.960

0.931

2

+ Actinobacteria

0.970

0.964

3

+ Deferribacteres

0.972

0.964

(B) Genus

# markers

marker

AUCtr

AUCval

0

class ~ sex + age

0.599

0.545

1

+ Lachnospiraceae NK4A136 group

0.907

0.852

2

+ Ruminococcaceae UCG-013

0.947

0.858

3

+ Bacteroides

0.955

0.915

Back to article page