Molecular Methods for the Detection of Waterborne Pathogens - Waterborne Pathogens: Detection Methods and Applications - page 225

Figure 8.1 Examples of FAME community profiles from groundwater samples. Com-
munity profile from well 63D. (a) Sampled June 20, 1996: (1) 14:0, (2) 15:0 Anteiso, (3)
16:1 ω 7c/15 Iso 2OH, (4) 16:1 ω 5c, (5) 16:0, (6) 17:0 Anteiso, (7) 17:1 ω 8c, (8) 18:3 ω 6c, (9)
18:2 ω 6,9c/18:0 Anteiso, (10) 18:1 ω 9c, (11) 18:1 ω 7c/ ω 9t/ ω 12t, (12) 18:0, (13) 20:0. (b)
Sampled August 15, 1996: (1) 12:0, (2) 13.566 (unknown), (3) 14:0, (4) 15:0 Anteiso, (5)
16:1 ω 7c/15 Iso 2OH, (6) 16:1 ω 5c, (7) 16:0, (8) 17:0 Anteiso, (9) 17:1 ω 8c, (10) 17:0, (11)
18:3 ω 6c, (12) 18:1 ω 9c, (13) 18:1 ω 7c/ ω 9t/ ω 12t, (14) 18:0, (15) 19:0, (16) 20:0. Source:
Adapted from Ref. 18 .
The main disadvantage of FAME analysis for detection and identification
of individual strains lies in the power of resolution of the method. Analysis
is limited due to the restricted number of strains and FAME profiles avail-
able in the BAME@LMG database. Currently, only about 50% of published
strains have standard FAME profiles. 16 To circumvent this problem, several
recent studies have incorporated machine learning techniques to allow bac-
terial identification at the species level using FAME data. Machine learning