Skip to main content

Table 2 Random forest classification of 25 freshwater samples with different level of fecal contamination

From: Fecal source identification using random forest

Random forest classifications
Environmental sample ID Type of sample Major type of contamination Level of fecal indicator bacteria Level of qPCR human marker‡‡ Clostridiales Bacteroidales
FMRMN73_092 Stormwater HC High High Sewagea (98) Sewagea (99)
FMRMN73_29 Stormwater HC High High Sewagea (84) Sewagea (89)
FMRHC33_42 Stormwater HC High Medium Sewagec (100)
FMRMN60_100 Stormwater HC High High
FMRMN29_108 Stormwater HC High High Sewagea (91) Sewagea (95)
MKE_162 River HC High Medium Sewageb (85) Sewagea (98)
MNE_163 River HC Medium Medium Sewageb (57) Sewageb (86)
KK_160 River HC Medium High Sewagec (77) Sewagea (99)
MNE_159 River HC Medium Medium Sewagec (68) Sewageb (98)
MKE_158 River HC Medium Medium Sewageb (98)
Gap_51 Harbor HC Medium High Sewagea (82) Sewagea (97)
Junction_54 Harbor HC Low Medium Sewageb (78)
Gap_55 Harbor HC Low Medium Sewagea (55) Sewagec (94)
Junction_52 Harbor HC Low Medium Sewagec (64)
FMRMN53_26 Stormwater NHC High Inconclusive
SHC12A_10 Stormwater NHC High Inconclusive Sewagec (90)
SMN17A_20 Stormwater NHC High Inconclusive Sewagec (100)
FMRHC43_43 Stormwater NHC High Not detected
FMRHAC22_38 Stormwater NHC Medium Not detected
Gap_53 Harbor NHC Low Not detected Sewageb (99)
1_mile Lake NC Not detected Not tested
2_miles Lake NC Not detected Not tested
DocIn_155 Lake NC Not detected Not tested
DocMid_156 Lake NC Not detected Not tested
DocOut_157 Lake NC Not detected Not tested
  1. HC human contamination (fecal indicator bacteria and human marker detected), NHC non-human contamination (fecal indicator detected and human markers not detected or inconclusive reflecting potential for low levels of human contamination), NC not fecal contaminated (fecal indicator not detected)
  2. Values in parentheses represent the proportion of sequences that belong to a given classifier among the total number of sequences from all classifiers
  3. Density levels of the fecal indicator E. coli and enterococci: not detected, 0; low, > 0–250; medium, 250–1000; high, > 1000 CFU/100 mL
  4. ‡‡Quantification levels of the markers human Bacteroides, Lachno2, and Lachno3 when tested: Not detected, 0; not quantifiable, > 0–15; low, > 15–100; medium, 100–10,000; high, > 10,000 gene copies/100 mL. In case of divergence between the human Bacteroides, Lachno2, and/or Lachno3 human markers, results were considered to be inconclusive. See Additional file 1 for details
  5. aIndex representing the percentage of the vote by the trees higher than the majority (50%)
  6. bIndex representing the percentage of the vote by the trees between 45 and 50%
  7. cIndex representing the percentage of the vote by the trees between 40 and 45%