Skip to main content

Table 2 Top species included in the GA selected 1, 2, 3, 4, 5 or 6 PCs produced with different data sets

From: Selection of microbial biomarkers with genetic algorithm and principal component analysis

Dataset for creating PCA

High contribution variables (high coefficients in the corresponding PC) included in the most selected components

Comp1

Comp2

Comp3

Comp4

Comp5

Comp6

Whole (PC1, PC7, PC2, PC27, PC11, PC15)

Prausnitzii–a

Gnavus+

Eutactus+a

Moorei–

Eggerthii–a

Zeae+a

Eutactus–a

Faecis–a

Prausnitzii+a

Obeum–

Dispar–a

Gnavus–

Formicigenerans–a

Copri+

Aerofaciens–

Lenta+a

Adolescentis+

Stutzeri+a

Catus–a

Muciniphila–a

Catus–

Animalis–

Mucilaginosa–a

Bromii+a

Faecis–a

Adolescentis–a

Adolescentis–`

Torques–

Aerofaciens+

Fragilis+a

Obesity (PC14, PC18, PC2, PC4, PC19, PC16)

Eutactus–a

Uniformis+

Dolichum–

Producta–

Caccae+a

Formicigenerans+a

Bromii+

Catus–a

Lenta–

Prausnitzii+a

Parainfluenzae+a

Bromii–

Adolescents–a

Dispar+

Aerofaciens+a

Aerofaciens–

Formicigenerans+a

Distasonis–

Formicigenerans+

Faecis+

Producta–

Fragilis–

Adolescentis–

Eutactus+a

Producta–a

Distasonis–a

Gnavus–

Faecis+a

Dispar–

Perfringens+a

Healthy (PC1, PC34, PC23, PC28, PC3, PC5)

Prausnitzii–a

Stutzeri–a

Callidus–a

Ovatus–

Copri+

Copri+a

Eutactus–a

Zeae+

Moorei+

Longum+a

Muciniphila–a

Muciniphila+a

Catus–a

Gnavus+

Formigenes+

Distasonis+a

Formigenes–a

Prausnitzii–

Formicigenerans–a

Dispar+

Prausnitzii+

Fragilis–

Catus+

Formigenes+a

Faecis–a

Lenta–a

Catus–a

Aerofaciens–

Biforme+

Eutactus+a

  1. Comp1, Comp2, Comp3, Comp4, Comp5 and Comp6 represent the 6 PCs selected by GA. For experiment with whole dataset they are PC1, PC7, PC2, PC27, PC11 and PC15 respectively; for experiment with obesity sample, they are PC14, PC18, PC2, PC4, PC19 and PC16; for experiment with healthy sample, they are PC1, PC34, PC23, PC28, PC3 and PC5
  2. aSpecies has a positive correlation with the probability of having healthy body mass
  3. + Positive correlation with the corresponding PC
  4. – Negative correlation with the corresponding PC