Scribd
Upload
118 views6 pages

Appendix D Sample BST Printouts and Results: Contract # 582-0-80121/ Work Order # 582-0-80121-08 - Final Report

The document is a final report analyzing bacteria source tracking results from various water and sediment samples. It includes summaries of classification results using a library of isolates to identify bacteria as being from human, bird, cow, dog, horse, or bat sources. Classification accuracy was highest when using a combination of antibiotic resistance profiles and carbon source utilization data, correctly identifying 94.7% of human/sewage isolates. Unknown samples were then classified using this library, with most isolates identified as being of human or bird origin, though classifications varied between samples.

Uploaded by

Malim Muhammad Siregar
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
118 views6 pages

Appendix D Sample BST Printouts and Results: Contract # 582-0-80121/ Work Order # 582-0-80121-08 - Final Report

The document is a final report analyzing bacteria source tracking results from various water and sediment samples. It includes summaries of classification results using a library of isolates to identify bacteria as being from human, bird, cow, dog, horse, or bat sources. Classification accuracy was highest when using a combination of antibiotic resistance profiles and carbon source utilization data, correctly identifying 94.7% of human/sewage isolates. Unknown samples were then classified using this library, with most isolates identified as being of human or bird origin, though classifications varied between samples.

Uploaded by

Malim Muhammad Siregar
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 6

Contract # 582-0-80121/ Work Order # 582-0-80121-08 Final Report

APPENDIX D SAMPLE BST PRINTOUTS AND RESULTS

Contract # 582-0-80121/ Work Order # 582-0-80121-08 Final Report

APPENDIX D BACTERIA SOURCE TRACKING RESULTS USING 6-WAY ANALYSIS (bird vs. cow vs. dog vs. horse vs. bat vs. human)

LIBRARY ISOLATES Table D-1 Summary Table Of Initial Library (2004) - Percent ARCC And RCC For Six-Way Classifications (Bird Vs. Cow Vs. Dog Vs. Horse Vs. Bat Vs. Human).

ARCC Bird Cow Dog Horse Bat Human Human-S Human-P Cross-val. ARCC

Human vs. Nonhuman All WilksAntibiotic lambda (stepwise) 38.1 34.2 45.6 46.2 37.9 38.4 37.9 35.2 48.5 42.0 34.8 20.1 25.2 25.2

Human S vs. Human P vs. Non-human All WilksAntibiotic lambda (stepwise) 36.8 31.6 45.0 43.3 34.2 34.2 37.4 34.6 47.0 38.5 31.9 16.2 16.0 39.6 33.7 11.3 39.6 30.3

Human P only vs. Non-human All WilksAntibiotic lambda (stepwise) 36.5 39.9 41.5 45.6 38.8 36.5 35.2 37.4 36.5 45.0 30.4 35.3

34.9

33.4

41.1 37.3

37.9 35.7

Notes: 1. Human-P = human isolates from portable toilets only 2. Human-S=isolates from sewage line only. 3. All antibiotic indicates the analysis was conducted using all the antibiotics. 4. Wilks-lambda (Stepwise) indicates that a stepwise method was used in which the program may remove some antibiotics from the analysis if they do not contribute to the classifications. 5. Abbreviations: ARCC Average Rate of Correct Classification; RCC Rate of correct classification

Contract # 582-0-80121/ Work Order # 582-0-80121-08 Final Report

Table D-2 Comparison of Discriminant Analysis Results Of The Revised Library Database Using ARP, CSU And A Combination Of All Data For Six-Way Analysis, Between The Human/Sewage Isolates Collected Only In 2005 And The Human/Sewage Isolates Collected in Both 2004 and 2005.

% Correct classification ARCC Human Non human Bird Cow Dog Horse Bat Cross-val. ARCC

ARP only HS 2004 HS 2005 and 2005 only 40.2 44.1 35.2 52.0

CSU only HS 2004 HS 2005 and 2005 only 62.4 67.2 71.2 93.0

ARP+CSU HS 2004 HS 2005 and 2005 only 67.9 72.2 74.9 94.7

44.3 40.4 40.0 50.8 35.6 38.0

45.5 41.3 41.2 49.7 35.6 41.7

59.3 62.4 59.4 56.5 56.5 52.7

61.7 68.5 62.9 57.0 61.8 55.7

65.3 70.0 68.2 62.2 59.7 56.5

65.3 74.6 72.4 63.2 64.4 59.9

Abbreviations: ARP antibiotic resistance profiles, CSU carbon source utilization, HS Human/Sewage

Note: Using the ARP+CSU combined database with the 2005 Human/Sewage isolates only, 94.7% of the human/sewage isolates were classified correctly. Use of the CSU data alone also provided a high level of correct classification for human; however there was a lower level of correct classification for the animals than the combined database. Results of cross-validation (also known as resubstitution analysis or leave-one-out method) showed only a slightly reduced ARCC for the ARP while for the CSU the difference was greater. This is partially an effect of the number of variables in the database (95 for CSU, 115 for combined ARP/CSU cf. 20 for ARP).

Contract # 582-0-80121/ Work Order # 582-0-80121-08 Final Report

UNKNOWN SAMPLE ISOLATES

Unknown source sample isolates were analyzed using by six-way classification using the 2005 human/sewage ARA + CSU library as this provided the highest RCCs for the six-way classification. Six-way classifications found relatively similar proportions of human source isolates as the two-way analysis. (Table D-3; Figures D-1 to D-18) Isolates identified as other animal sources varied by sample. It should be noted that Discriminant Analysis identifies a source for each isolate so isolates from sources other than those included in the library will identify as one of those in the library. For example, the lack of wildlife in the library will result in any wildlife isolates being classified perhaps as bird or bat. There were a limited number of isolates for each water/sediment sample so small differences in percent may only reflect a difference of one or two isolates.

Contract # 582-0-80121/ Work Order # 582-0-80121-08 Final Report

Table D-3 Six-way percent classification of unknown source isolates from each sample as Human (Sewage), Bird, Bat, Cow, Horse or Dog identified using the library including the 2005 sewage only. Date received Sample ID Matrix Human Bird / Sewage 17.4 23.9 24.1 24.1 15 35 18.8 21.9 5.3 42.1 16.3 20.4 18.2 13.6 20.7 6.9 43.3 3.3 31.9 10.6 33.3 6.1 58.8 3.9 59.1 2 42.9 21.4 41.5 17 77.6 10.2 52.6 8.8 74.5 2.1 %Isolates Cow Horse

Dog

Bat

8/3/2004 (Dry Weather)

1 2 3 7

3/16/2005 (Dry Weather)

4 5 9 10 7 8 9 10 11

4/26/2005 (Wet Weather) 5/17/2005 (Dry Weather) 7/13/2005 (Wet Weather) 7/14/2005 (Wet Weather)

Water Sediment Water Sediment Water Sediment Water Sediment Water Water Sediment Water Water Water Water Water Water Sediment

6.5 3.4 15 6.3 5.3 2 18.2 6.9 18.3 10.6 3 13.7 8.2 7.1 3.8 0 3.5 0

10.9 20.7 2.5 18.8 21.1 8.2 13.6 20.7 30 21.3 21.2 5.9 12.2 8.9 18.9 2 8.8 17

37 10.3 27.5 3.1 15.8 28.6 18.2 37.9 5 10.6 15.2 7.8 10.2 1.8 9.4 4.1 15.8 2.1

4.3 17.2 5 31.3 10.5 24.5 18.2 6.9 0 14.9 21.2 9.8 8.2 17.9 9.4 6.1 10.5 4.3

Water

51.9

7.7

15.4

15.4

3.8

5.8

11

Water

67.3

7.7

3.8

11.5

1.9

5.8

*as described on chain of custody forms received from UH with the samples

You might also like