Journal of General Microbiology (1986), 132, 1827-1842.
Printed in Great Britain
1827
A Probability Matrix for the Identification of Gram-negative, Aerobic,
Non-fermentative Bacteria that Grow on Nutrient Agar
ByBARRY HOLMES,* CLA IR E A. P I N N I N G 7 A N D
C H R I S T I N E A . DAWSON
Computer Identfication Laboratory, National Collection of Type Cultures,
Central Public Health Laboratory, Colindale Avenue, London NW9 5HT, UK
(Received I5 November I985 ;revised I3 February 1986)
Results of the identificationof 621 strains of Gram-negative, aerobic, non-fermentative bacteria
by a computer-based probabilistic method are given. Although many of the strains were atypical
and have caused difficulty in identification in the medical diagnostic laboratory, the
identification rate on this matrix was 91.5%.
INTRODUCTION
Dybowski & Franklin (1968) were the first to describe the use of a computer-assisted,
conditional probability method for the identification of enterobacteria. Lapage et al. (1970)
successfully identified 70 to 80% of 279 freshly isolated strains examined in both a limited and a
more extensive number of tests. Lapage et al. (1973) discussed the general aspects of
probabilistic identification; the mathematical model used was described by Willcox et al.
(1973). Bascomb et al. (1973) published a matrix for the identification of Gram-negative rods of
clinical importance and presented the results obtained in the identification of 1079 reference
strains. The matrix contained 56 fermentative and 14 non-fermentative taxa and, although the
method worked well for the former group of organisms, it was less successful for the latter group
due to the unsuitability of certain tests in the matrix. A new matrix was constructed over the
ensuing years and used as the basis of an identification service provided by our laboratory.
The methods employed in the operation of the identification service were described by
Willcox et al. (1980). In the operation of the identification service, the test results obtained for
strains submitted for identification were accumulated by computer. These results were then
sorted by taxon and printed in the form of summaries as described by Holmes & Hill (1985).
From these summarized results, the matrix for the non-fermentative organisms has been
developed in its present form. Following evaluation, this matrix is now in current use for the
routine identification service.
In this paper are presented the results obtained in the identification of 621 strains of bacteria
belonging to 66 taxa in the present version of the probability matrix.
METHODS
Overallprocedure. A matrix was constructed which comprised the probability of a strain of any given taxon
yielding a positive result in each of the chosen tests (Table 1). Individual strains were then identified on the basis of
these results.
Tam. The taxa incorporated in the matrix are given in Table 1. In the oxidation/fermentation (O/F test) of
Hugh & Leifson (1953) the organisms gave either an oxidative result, produced no acid in the test or developed an
alkaline reaction. These organisms are referred to as non-fermentative bacteria.
The taxa were selected primarily to include those of known medical importance and those likely to occur in
Present address: Gibco-Sensititre, Imberhorne Lane, East Grinstead, West Sussex RH 19 lQX, UK.
0001-3080 0 1986 SGM
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Wed, 12 Jul 2017 16:26:54
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Wed, 12 Jul 2017 16:26:54
I 1 E 9 t I 6 6 6 6 1 I I I I I I I I I I 6 6 6 6 6 6 1 I 0 1 6 6 1 99OS6666
ssv lol!uuw -0s
I 6 6 P 6 Z 6 1 66661 I I I I I I O E I I 6 6 6 6 6 6 6 6 1 6 I 6 6 6 6 1 666666
SSV W P M '69
I 6 6 S S I I I I I I I I I I I I I I086666661 OLI I I I 6 6 E l I
SSV =013a? '8P
I I I I I E6 6 61 I I I I I I I I I I 6 6 6 6 6 6 1 I I I 6 6 6 6 6 6 6 6 6 6
ssv lol!souI 'LP
ssv loJa349 'W
I 96 9L If I 66 66 I I I 8 I I I I I I 09 66 66 26 I I I 66 66 66 66 66 66
I I I I I I I I I I I I I I I I I I Z61 6 6 1 I I 6 6 1 6 6 6 6 1 P6
ssv i w i n a *SP
I 66 8 I I P I I I I I I I I I I I I 66 66 66 66 I L8 I 66 66 I 66 66 66
SSV aso!qoIl~3 *W
I 66 Z I I 66 66 I I I SZ I I I 8E I I 66 66 66 66 W 66 E 66 66 66 66 66 66
SSV =Ou!qOJV 'EP
I I I I I 1 1 1 1 I I 8 I I I I I I I 6 6 6 6 6 6 1 I I 6 6 1 I 6 6 1 66
ssv lol!uoPv 'ZP
ssv =oonl!I 'It
I 6 6 9 8 S 8 I 66661 I I I I I I 8 Z I 166666666LP6626666666666666
= V N a '09
6 6 1 9 8 6 6 L I OS681 I 6 Z l OZI 1 8 6 s I OZI 1 I I P I I I I I I I
5 6 1 6 6 9 6 1 I E E I W 8 6 1 I I I P6 1 I I I I I I I P I I I I 1 I I
'PAY U!aS83 '6E
I I I 1 l l I 1 I I I I l l I I E I l I l l l l l l l l l l
al!ualaS %kO '8E
I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I
(s@aso3nlsSMd'LE
I 11 I 8 I I EE I I I I 51 I I I I I I I I I I 66 I I I I I I I
(P!3e)aso3nI%SMd'9E
1 1 1 1 6 6 1 1 1 I I I I I 1 Z 8 1 1 I I I I I I I 1 I I I I I
-3ap au!ql!uJo 'SE
I I 1 1 6 6 1 1 I I I I I I I I I I I I I I I I I I I I I I I
w p au!sLT -PE
I 1 1 I I I 6 6 1 I I S ZI 9 I I I I I 6 Z L I I I I I I 6 6 E E I L ZZ
' y p au!u!%Jy 'EC
1 1 I ~ 1 l 1 1 1 1 1 8 I l 1 1 Z l I I l l Z l l l l l l l l
Vdd 'ZE
1 6 6 8 6 2 1 1 I 6 6 1 I I I Z I I 1 I I oP9P66661 I I I 6 6 1 66669
9 d N O 'IE
I I I I I W661 I I I I I E Z WS I I I I S OSI I I I I I 9
a1eqew 'OE
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
aieuoonI9 yjz
I I l l l I I I l l I l l l L 6 l I I I 1 l l l l l l l l l I
ISI S'H '8Z
I I 8 I I I I I I I 8 I I I L 6 S L I 9 6 1 I I I I I I I 6 6 1 0s
Jaded SzH 'LZ
L P 2 P I P I E 1 6 6 1 I I I I Z6 ZI OLS 6 I I I I S ESE81 I I I 1 0 s
N3X '92
aiqd u!ieiag 'sz
66PI66961 98661 66661 I I I 6 6 L L I I 1 I I Z L Z I I I I I I 9
q w ullala9 'PZ
861 86281 1 1 1 1 9E961 I I 1 9 8 s Z I I I I 8 S 1 I I I I I I I
asaaJn 'EZ
66 96 EE P I W 95 66 66 I 66 I 66 66 ZE I I I 66 66 ZP L l ZP I 66 66 66 66 66 66
'J1!3 TIalS!Jqa 'ZZ
I I LP I I 66 66 I I I I E E9 L6 EZ S6 E6 W I2 0s €8 Of P8 66 I 66 66 66 66 68
+J1!3 SUOUrUr!s 'IZ
I I 81 I I 6 6 6 6 1 I I I I I L66 E8LLI IZ I 8 9 E P 8 E 6 I 0s 1 I L P I I
3npaJ all?Jl!N 'OZ
I I P P 6 6 1 €9661 L 6 6 I S I 66E6I 6609621 I 8 P 86666666666666
6 6 1 I P I I I SZI 8 6 0 S I 6 6 L 6 1 8 6 6 8 6 1 I I I 9 8 1 W l I I I I I
au!le?le 7 aP H '61
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
*wJaJ7 aP H '81
I 66 26 ZP I 66 66 I I 1 I S I I 21 I I 08 EE 66 66 I 66 I I 0s 66 0s E6 P6
'P!XO 7 '4p H 'LI
66 I 8 85 66 I I 66 66 66 66 56 66 66 88 66 66 OZ L9 I I 66 I 66 66 0s I 0s L 9
UOY!a? 0 @nH '91
aseP!xo 'S 1
66 66 66 66 66 I I 66 66 66 26 66 I 66 66 56 E6 66 66 66 66 I I 66 66 66 66 66 66 66
asalala3 '91
66 66 66 66 I 66 66 66 66 66 66 66 66 66 66 66 66 66 66 66 66 66 66 86 66 66 66 66 66 66
6 6 6 6 6 6 5 8 I 6 6 6 6 6 6 1 S6 S LLZOS 6 6 6 6 6 6 6 6 OZ6 6L9 1 9 6 6 6 6 6 1 6 6 6 6 6 6 6 6 6 6
AaquoD3epJ 'E I
lua~%!da~ueJO'ZI
I I I I I I I I I I I I I I S8I I I I I I I I I I I I I I I
iuaux8!d pay I
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 I I I I I l I I I
I I I I I I I I I I I
luawg!dMollaA '01
1 1 1 1 1 1 1 1 1 1 1
luaw8!d UaaJD '6
1 1 1 1 1 1 1 1 1 1 1
iUawS!d lalO!A '8
1 1 1 1 1 1 1 1 1 1 1 1
1uawS!d UMOJB ' L
iuaurS!d Jwa.
86
1
1
1
66 SZ 96 I
1 1 1 1
1 1 1 1
1 1 1 1
66 66 I I I
1 1 1 1 1
1 1 1 1 1
1 1 1 1 1 1
I
1
1
1
E
1
1
1
I
1
1
8
I
1
1
1
I
1
1
0
I
1
1
1
I
1
1
1
66 I
1 1
1 1
1 1
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
Z
66
66
I
I
I
66
66
I
SL
66
66
I
1 I
P 66 I I 66 66
I 66 66 66 58
66 66 66 66 66
16666661
I 98 66 66 I
66
66
I
I
66
66
66
I
I
66
E8
66
I
I
L6 61
19 OS
66 66
1 I
I I
66 S 66 66 I 66 66
66 66 66 66 66 66 66
66 16 66 66 I LP I
L6S6980609961
Z6 6L P8 06 I P I
66
66
I
I
I
66
66
66
I
I
66 66 66 66 66 66 66 66
66 66 66 66 66 66 66 66
66 16 66 66 66 66 66 66
IS8666666666666
I 6L 66 66 I 0s 66 95
aAhpeSau luaus!d
(La)WMOJf)
'9
'5
'P
(DoLEI WMoJD 'C
(LH)4!1!10N 'Z
(DOLE)Al!I!lWI ' I
OE 62 8Z LZ 9Z SZ PZ CZ ZZ IZ OZ 61 81 LI 91 SI P I E l Zl I 1 01 6 8 L 9 S P E Z I
N O S M V a 'V ' 3 a N V ONINNId 'V '3 'SBMV?OH
'a
8Z8 T
Probabilistic identification of non-fermentem
1829
Table 1 (continued)
Character 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66
1. 1 1 1 1 1 1 1 1 1 1 1 1 1708899368859 199 16499 21388997494 16370 1 127
2. 1 1 199 1 1 1 1 1 1 1 1 19589999099999999 19999642291999995999980 16082
3. 98 99 99 1 99 99 99 99 99 99 99 99 99 99 99 99 99 99 90 99 99 99 99 99 99 99 99 99 99 99 67 99 99 70 10 64
4. 99 99 99 99 1 99 20 99 99 99 96 99 99 99 99 99 99 99 99 99 99 89 99 99 99 99 99 99 99 99 99 99 99 99 99 99
5. 1 14 56 50 99 99 99 99 99 99 99 99 99 99 43 99 91 94 99 99 99 99 95 40 1 99 96 99 99 54 99 99 1 99 1 9
6. 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
7. 1 1 1 1 1 1 1 1 1 1 1 1 1 1 3 3 1 1 6 1 1 1 1 1 1 1 1 4 1 1 1 0 1 1 1 1 1 1
8. 1 1 1 5 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
9. 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 3 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
10. 998643 1 1 1 1 1 1 1 1 1 1 1 1 1 8 1 1 1 1 1 4 6 0 9 9 1 1 1 1 3 2 1 1 1 1 9 9 9 1
11. 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
12. 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 4 1 1 9 9 1 1 1
13. 93 99 99 87 1 1 1 1 90 99 54 99 99 95 99 99 99 99 99 99 99 1 99 99 4 93 99 99 99 99 99 99 30 99 80 99
14. 99 99 99 99 1 1 1 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99
15. 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 93 99 99 99 99 1 1 32 99 99 99 99 99 99 99 99 99 99 99 99 73
16. 25 1 1 6 99 99 99 99 99 99 96 1 99 99 2 99 1 99 1 1 99 1 69 10 65 3 99 1 1 17 1 99 99 10 30 18
17. 75999994 1 1 1 1 1 1 4 9 9 1 198 199 19999 19931903597 199998399 1 1907082
18. 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
19. 2 1 1 6 1 1 1 1 7 0 1 2 4 19988 199 199 1 1 1 1 3 9 1 1 185 1 1 1 1 9 9 1 1 1 1
20. 30 1 99 94 99 1 1 1 40 96 58 99 1 97 99 99 24 13 24 1 1 99 59 99 1 66 99 99 13 99 1 38 1 1 10 1
21. 4 1 1 99 1 1 1 1 1 1 1 1 36 91 99 99 99 1 94 85 1 1 2 99 4 98 41 99 94 96 99 99 1 1 1 82
22. 49 1 1 99 1 1 1 1 1 1 13 1 57 97 99 99 99 1 98 99 1 22 93 99 28 99 36 99 98 92 99 99 1 1 99 99
23. 19 99 99 13 1 1 1 99 1 84 4 1 1 32 75 1 54 6 38 8 1 33 6 99 1 99 16 33 26 33 67 46 1 99 1 1
24. 98 1 1 1 1 1 1 9 9 1 1 4 1 1 172 1516348 1 1 189 1 1 3 9 196 1 1 1 1 1 1 I 8 2
25. 99 79 86 88 1 1 1 99 1 1 92 1 1 6 86 1 67 87 58 1 1 78 97 1 8 51 2 99 1 1 1 1 80 1 1 91
26. 7 7 1 6 1 1 1 1 1 1 8 1 1 1 3 6 4 121191623 1 1 5 2 1 0 1 2 19916 5 117 1 1 1 1
27. 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 1 1 1 1 1 1 1 1 4 1 7 0 2 124 1 1 1 1 1 1 1 1 1
28. 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
29. 1 1 1 1 1 1 1 1 1 1 1 1 1 173 1 7 1 1 5 1 1 1 1 1 1 1 1 1 1 1 1 3 3 1 1 1 1 1
30. 2 1 1 5 3 1 1 1 1 1 1 1 1 14091 117 148 1 1 15599 1 1 1134358 121 1 1 1 1
31. 26999913 1 1 1 1 1 1 1 1 1 1 1 1 139 1 1 1 1 1 1 3 199 1 1 1 1 1 1 11090 136
32. 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 40 1 1 1 1 1 11 I 1 1 1 1 1
33. 1 1 1 1 1 1 1 1 1 1 1 1 1 1 9 8 4 0 1 1 9 8 9 9 199 199 1 1 4 0 9 9 9 9 4 9 9 9 1 1 1 1 1
34. 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 163 1 1 1 1 182 1 1 1 1 1 1 1 1 1 1 1 1 1
35. 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 4 6 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
36. 28686 1 1 9 9 1 1 1 1 199 1 1 2 1 3 3 13499 1 3 3 1 1 1 1 19639 5 3 3 1 1 1 1 1
37. 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
38. 1 1 1 1 1 1 1 1 1 1 1 1 1 1 7 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1
39. 99 1 1 7 0 1 1 1 1 1 13899 1 1 8 6 195 165 1 1 197 1 1 1 1 9 9 1 1 1 1 1 13082
40. 869986 1 1 1 1 1 1 1 1 1 1 5 2 1 1 1 7 2 1 1 1982082 1 1 6 1 1 1 1 1 1 8 0 9 9
41. 99 99 99 99 1 1 1 1 1 1 1 99 1 15 99 1 99 199 99 199 53 99 99 99 18 99 99 99 99 199 99 99 99
42. 1 1 9 9 3 1 1 1 1 1 1 1 1 1 1 1 1 199 1 3 1 1 1 1 1 1 1 1 31810 1 1 1 1 9 9 1 1
43. 23219999 1 1 1 1 1 1 1 1 1 189 199 18799 189 3 19999 399991299 130996055
44. 5999999 1 1 1 1 1 1 199 1 1 1 1 9 9 11899 1 3 3 9 19991 19912 199 180996099
45. 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 9 9 1 1 1 1 1 1 1 1 1 1 9 9 1 1 1 1 1 9 9 1 1
46. 61999999 1 1 1 1 1 1 1 9 9 19199 199 18492 189 1 1 1 8 9 5 3 9 9 9 8 2 8 2 9 9 4 6 199 1 9
,47. 1 1 199 1 1 1 1 1 1 1 1 1 3 1 1 9 9 1 5 3 8 199 1 1 1 1 199 1 2 9 9 1 1 9 9 1 1
48. 19999 1 1 1 1 1 1 1 1 1 1 1 1 199 1 1 9 2 1 1 1 1 1 9 9 8 8 199 1 1 6 7 1 1 9 9 1 0 2 7
49. 96999999 1 1 1 1 1 1 199 1 1 1 1 9 9 1 1 9 9 13395 19989 199 37199 199999982
50. 599 185 1 1 1 1 1 1 1 1 19186 199 17523 167 1 1 1 4 4 199 64299 1 199 1 1
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Wed, 12 Jul 2017 16:26:54
1830
B. HOLMES, C . A . P I N N I N G A N D C . A . D A W S O N
Table 1 (continued)
Character
1 2 3 4 5 6 7 8 9 10 1 1 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
51. Raffinose ASS
52. Rhamnose ASS
53. Salicin ASS
54. Sorbitol ASS
55. Sucrose ASS
56. Trehalose ASS
57. Xylose ASS
58. Ethanol ASS
59. Fructose ASS
60. Thornley Arg
61. Tween 20
62. Tween 80
63. Tyrosine hyd.
64. Tyrosine pig.
65. Nitrite reduc.
66. PHBA, growth
67. PHBA inclusions
68. Aesculin
69. Cetrimide
70. Fluor. King's B
71. Growth (5 "C)
72. Growth (42 "C)
73. 3-Ketolactose
74. Lecithinase
75. Starch hyd.
76. 10% Glucose
77. 10% Lactose
78. 10% Bile, growth
79. 40% Bile, growth
80. Bile stimulation
8 1. Haemolysis
82. 4% NaCI, growth
83. Ehrlich indole
1 1 1 1 1 9 9 1 1 199-509980 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 9 9 1
9999999999 1 169 192999999 1 1 1 1 1 1 1 1 1 18999 1 4 146 1
179950 1 5 0 9 9 1 1 1 9 9 9 9 9 9 9 9 1 1 1 1 1 1 1 1 1 1 7 5 1 1 1 1 9 9 1
679999 1 199 1 1 1999999 1 1 1 1 1 1 1 1 1 1 1 1 1 9 3 1 1 1 1 1
999999 19999 1 1 199999980 1 125 1 1 2 1 1 1 11121 112 199 1
839999 1 9 9 9 9 1 1 1 9 9 9 9 9 9 9 9 1 1 1 1 1 1 1 1 1 1 9 9 9 9 1 1 8 6 9 9 1
99 99 99 99 99 99 99 99 50 99 99 99 99 1 1 1 1 1 1 1 1 1 1 99 99 1 4 6 99 1
99 93 99 99 1 99 46 88 76 1 17 99 1 33 85 1 1 1 1 8 1 1 1 99 99 1 27 57 1 1
99 99 99 99 99 99 16 10 9 99 99 99 60 1 3 1 1 1 2 1 1 1 1 99 99 1 8 57 99 1
113 1 1 9 9 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 9 9 1 1 1 1 1 1
1 67 1 1 50 99 1 99 97 1 1 1 99 11 5 99 4 1 37 33 99 69 1 99 93 77 99 98 99 99
1 1 1 1 1 1 19990 1 1 1 1 5 199 1 1 1 1 2 2 1 1 5 6 1 1 3 9 6 9 9 6 8 8
1 67 99 1 99 1 99 83 23 8 1 4 99 92 95 50 99 1 1 8 18 1 1 99 1 1 8 16 1 56
113 1 1 1 1 1 7 6 1 1 160 3 254 19910 866 1 12236 1 8 4 9 419
9999 1 9 9 1 1 5 1 1 1 1 1 4 6 1 1 9 7 4 5 4 1 1 3 2 5 1 1 9 9 1 1 1 4 3 7 1 9 8
99 99 99 99 99 99 99 97 89 99 99 99 99 98 98 90 99 1 7 1 91 1 50 99 99 1 99 99 96 95
180 1 1 1 1 9 8 2 0 792 1 1997953 176 1 1 1 2 150 1 1 1 1 1 1 1
199 1 19999 1 1 167999999 1 1 1 1 1 1 1 1 1 199 1 1319699 1
620 1 1 1 1 9 9 3 8 2 817 4 18788 6 1 1 1 1 1 1 13393 1 1 1 0 1 5
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
1 1 1 1 1 1 1 11 40 I 1 1 1 1 7 34 1 1 1 1 1 1 1 1 1 1 12 2 1 7
17 80 99 1 99 99 62 77 42 1 1 1 1 53 81 46 64 1 14 75 95 3 1 99 1 1 4 12 1 1
1 1 1 1 1 1 1 1 1 1 1 8 8 9 9 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
1 7 1 1 1 1 1 8 9 4 3 1 1 1 1 1 1 8 9 1 1 3 1 1 3 1 1 1 1 4 1 0 4 1
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 2 1 1 1 1 1 1 5 1 1 1
28475099 1 1 196 192 1 4 1 1 1 1 1 1 1 1 1 1 19999 1 1 2 7 9 6 1
1 1 1 1 1 1 1 9 4 19217 4 1 1 1 1 1 1 1 1 1 1 19999 1 1 896 1
0 0 0 0 0 0509994 0 0 0 05099 0 0 0336799 3 0 0 0 0 0 0 0 0
0 0 0 0 0 0 19994 0 0 0 05099 0 0 0254499 3 0 0 0 0 0 0 0 0
0 0 0 0 0 0 1 1 2 8 0 0 0 0 1 1 0 0 0 1 1 1 3 0 0 0 0 0 0 0 0
0 0 0 0 0 0 1 1 4 7 0 0 0 05050 0 0 121 15083 0 0 0 0 0 0 0 0
0 0 0 0 0 050 142 0 0 0 09950 0 0 142 150 1 0 0 0 0 0 0 0 0
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 119983 1 1 116049 1 1
medical specimens. In the constructionof the matrix, however, many taxa were found to be poorly defined, so that
taxonomic studieswere required to improve their classification in order to achieve satisfactoryidentification.The
results of these taxonomic studies have been published separately and will be mentioned under the results for
separate taxa. The majority of the taxa are recognized species or genera. Some are without formal names, such as
Group IIf (Tatum et ul., 1974).
Tests. The range of tests has been described by Holmes et al. (1975), with the following additions: growth and
stimulationof growth on 10% (w/v) and 40% (w/v) bile, haemolysisand growth on 4% (w/v) NaCl (incorporatedto
differentiate primarily between Branhumella, Moruxella and Neisseriu), and Ehrlich's indole (incorporated to aid
identificationof certain Flavobucterium species). For these six additional tests, probabilities are not allotted in the
matrix for all taxa. In practice, 83 test results were available; 73 tests were set up of which one, pigmentation, if
present, provided a choice of seven possible colours and another, the O/F test, four possible results. Except for
some taxa (in the case of the six tests mentioned above) the 83 test results were allotted probabilities for each taxon
in the matrix (Table 1).
Methods. Most of the methods used have been described previously (Holmes et al., 1975). Additional tests were
as follows. The three bile tests were done using a blood agar plate divided into three sections by two parallel cuts.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Wed, 12 Jul 2017 16:26:54
Probabilistic identijication of non-fermenters
1831
Table 1 (continued)
e
c-.
Char-
acter 31
32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66
51. 1999946 1 1 1 1 1 1 1 1 1 1 I 120 1 8 1 1 1 1 196 1 1 1 1 1 1 1 190 1 1
52. 2 79915 1 I 1 1 1 1 1 1 1 1 1 I 1 13592 1 1 1 1 1 6 1 1 1 1 5 199 13099 1 1
53. 29999 1 1 1 1 1 1 1 1 1 1 1 1 187 1 1 1 1 1 2 182 1 1 1 1 1 1 1 1 9 0 1 1
54. 1 1 199 1 I 1 1 I 1 1 1 1 1 1 1 9 9 136 1 1 7 8 1 1 1 1 199 1 1 1 1 199 1 1
55. 23999992 1 1 1 1 1 1 199 1 1 1 1 5 5 14062 1 1 1 1 199 1 118 1 1 1 1 1999999
56. 869999 I 1 1 1 1 1 1 1 1 1 147 199 16792 189 1 19917 199 1 1 1 1 1992099
57. 18999999 I 1 1 1 1 1 1 1 1 192 197 19999 1 1 2 8 0 9 9 9 9 147998499 1 199 191
58. 18 99 1 1 1 I 1 1 60 1 1 99 1 13 89 1 1 80 18 1 1 I I 1 99 92 1 86 24 64 82 33 1 90 20 1 1
59. 70 99 99 99 1 1 1 1 1 1 1 99 1 96 96 I 99 1 97 99 1 22 45 99 98 92 83 99 83 88 99 1 1 99 70 99
60. I 1 1 1 1 1 1 1 1 1 1 1 1 19720 1 19792 199 199 1 12799974099 1 1 1 1 1
61. 99 99 99 85 1 99 20 99 90 96 99 99 7 99 99 99 99 93 81 99 99 99 99 99 99 99 44 94 44 93 99 99 99 1 99 91
62. 93 99 99 77 1 50 1 1 10 80 67 1 1 98 94 99 99 1 53 8 1 67 99 99 98 99 13 71 10 90 1 99 20 1 90 99
63. 67 1 1 85 1 1 1 1 1 1 1 1 1 99 96 99 97 53 94 99 1 89 93 99 89 99 92 99 89 84 99 99 30 10 10 18
64. 86 1 1 1 1 1 I 1 1 1 1 1 1 35820 1992799 1 I909046 470 1804499 190 1 1 1
65. 25 1 1 125 1 1 1 1 1 1 199 195 1 1 1 9 1 1 1 3 9 9 1 4 8 1 1 9 9 1 1 1 1 1 1
66. 99 99 99 99 1 1 1 1 99 24 21 1 93 99 99 80 99 99 99 99 99 99 83 99 96 99 95 99 99 99 99 99 99 99 99 99
67. I 1 1 1 1 1 1 1 9 9 4 5 19397 1 1 8 9 9 9 5 19999 13090853599 5 1 1 9 9 9 9 9 9 6 0 1
68. 95999993 1 1 1 1 1 1 199 1 120 130 1 1 1 1 181 199 1 118 1 5 1 1 1999099
69. 5 1 1 8 1 1 1 1 I 1 1 1 125998083 19354 1 1 5 1 9 9 1 2 8 3 1 8 9 4 4 4 9 9 1 1 1 1 1
70. 1 1 1 I I 1 1 1 1 1 1 1 1 1 6 9 I 1 172 1 1 1 1 1 1 1 1 1 3 1 199 1 1 1 1 1
71. 1 1 1 199 1 I 1 1 116 1 1 110 1 1 7 67899 1 1 6 1 0 1 6 3 5 1405999 1 1 1 118
72. 26 1 99 1 1 1 1 99 1 1 25 1 99 3 99 99 36 1 1 1 1 67 29 99 1 56 99 99 4 92 1 8 10 1 1 1
73. 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 l l l l l l l l l l l l l l l l
74. 79 1 1 60 1 1 1 99 1 60 83 1 43 2 96 50 98 1 83 8 1 22 82 70 1 32 14 99 1 78 1 1 1 1 1 73
75.631 1 1 1 1 1 1 1 1 1 9 9 1 2 1 1 1 1 1 1 1 1 1 1 4 9 9 1 1 2 5 0 1 1 1 1 1 9 9
76. 129999 1 1 1 1 1 1 1 1 1 1 187 199 18299 1 1 199 186 199882533 1 1 1 1 1
77. 18699 1 I 1 1 1 1 1 1 1 1 1 3 5 192 15499 178 1 1 4 8 7 19960 199 1 130 1 1
78. 099 0 02550 1 17099799999 0 0 0 0 0 0 0 0 099 0 099 0 0 0 0 0 0 0 0 0 0
79. 099 0 0 1 1 1 I209954 199 0 0 0 0 0 0 0 0 099 0 099 0 0 0 0 0 0 0 0 0 0
80. 0 1 0 0 1 1 1 1 1 9 6 4 1 1 0 0 0 0 0 0 0 0 0 1 0 0 1 0 0 0 0 0 0 0 0 0 0
81. 0 1 0 0 1 1 6 0 9 9 1 5 3 3 1 1 0 0 0 0 0 0 0 0 0 5 0 0 0 5 0 0 0 0 0 0 0 0 0 0 0
82. 099 0 0 199 1 1 19988 199 1 0 0 0 0 0 0 0 0 1 0 099 0 0 0 0 0 0 0 0 0 0
8 3 . 9 5 1 1 1 1 9 9 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
The outside two sections of blood agar were removed and replaced by pouring 10% (w/v) bile blood agar into one
section and 40% (w/v) bile blood agar into the other, to the level of the original blood agar strip. The test strain was
inoculated across all three agar types in a single streak. Growth, and stimulated growth on either bile
concentration or on the central blood agar strip edges immediately adjacent to either bile strip, were read using the
centre of the blood agar strip as a control. Haemolysis was determined on blood agar and '4% NaCl growth' on
blood agar containing 4% (w/v) NaCl. These additional tests were read at 1,2 and 5 d, except for haemolysis which
was read at 24 h. For Ehrlich's indole, strains were inoculated into peptone water and incubated for 5 d. Xylene
(1 ml) was added to the culture, which was shaken vigorously.A few drops of Ehrlich's reagent were then run down
the side of the test tube. Indole production was indicated by the development of a red colour in the reagent phase.
Except where otherwise required by the specification of the test, incubation was at the optimum growth
temperature for the organism, usually 37 "C,but sometimes30 "C, and occasionally room temperature (18-22 "C).
Coding of tests. The methods were as described by Bascomb et al. (1973), except that the choice of pigment
colours was increased to seven (Table l), acid production from carbohydrates.and gas from glucose in
peptonelwaterlsugar medium were recorded as four-state (+, &, T , -), and the O/F test was also coded as fourstate (oxidative, fermentative, alkaline and negative, when the medium was unchanged). The fermentative state
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Wed, 12 Jul 2017 16:26:54
1832
B. HOLMES, C. A . PINNING AND C. A . DAWSON
was included in case a fermentative strain was mistakenly processed through this matrix - the fermentative result
would contribute to the non-identification of such a strain.
Linkage oftests. Test linkage was taken into account in the calculations for certain tests (Willcox et al., 1973).
Some tests were done at more than one temperature and allowance was made for growth at the different
temperatures [e.g. motility at room temperature (RT) or 37 "C].The two tests for gelatinase production - the stab
liquefaction and plate methods -were treated as linked. Failure to produce acid from glucose/peptone/waterwas
linked with failure to produce gas. Similarly, failure to grow on /I-hydroxybutyrate was linked with failure to
produce lipid inclusionsafter growth on that medium. The tests for H2Sproduction, using lead acetate papers and
triple sugar iron (TSI) agar, were also treated as linked. Although the biochemical pathways involved in these tests
are dissimilar (Lapage & Bascomb, 1968), there was a marked correlation between the results obtained, i.e. strains
which were positive in TSI were almost invariably positive by the lead acetate paper method. By linking the tests,
the recording of false positives in the TSI agar, due to pigment production, is avoided.
Pigment production with the seven possible colours was treated as a multistate test (Willcox et al., 1973). The
O/F test was coded in a non-standard way by the identification program. In calculating the identification scores,
the first component, with the name Hugh & Leifson (Table l), was treated as 'negative or alkaline', so the entries of
the first three components for a taxon should add up to a nominal 100%.The fourth component, H & L (Hugh &
Leifson) alkaline (Table 1) was treated as an independent test, so it could take any value less than or equal to the
value of the first component. Possible entries would then be: 01, 01,99,01 (all strains fermentative); 99,01,01,01
(all strains negative); 99,01,01,99 (all strains alkaline); and 99,01,01,25 (25% alkaline, 75% negative). In test
selection and printed output, however, the O/Ftest was treated as a four-state test. On the identification reports,
the percent probability displayed was for a positive result in a particular test. Therefore the figure displayed for
Hugh & Leifson for a negative result was the percent probability (%P)of Hugh & Leifson not negative and was
calculated as follows :
%P [H & L (negative or alkaline)]- %P [H & L (alkaline)]= %P [H & L (negative)]
%P [H& L (not negative)] = 100%- %P [H& L (negative)]
Adjustments needed to be made before and after this calculation to allow for matrix figures of 1 and 99 being
used in place of 0 and 100 respectively (see below).
Construction of matrix. The methods were similar to those of Bascomb et al. (1973). However published data,
including results in standard textbooks, were not used. The probability values allotted for each taxon in each test
had upper and lower limits of 0.99 and 0.01, and the probability value of the actual percentage of strains of each
taxon found to give a positive result in a particular test was used. The matrix is given in Table 1.
Probabilistic identijfcation. The methods were as described by Bascomb et al. (1973). In certain cases
identification to a combined taxon (or composite group; see Holmes & Hill, 1985) was permitted. A combined
taxon comprised two closely related taxa which had proved difficult to separate in the construction of the matrix,
as there were few,or no constant (0.99/0.01) characters which would differentiatebetween them (see also results on
individual taxa). The combined taxa were Alcaligenes denittificans and A .faecalis, Flavobacterium meningosepticum
and Flawbacterium species Group IIb, Kingella denitrijicans and K . kingae, Pseudomnasfluorescens and P . putida
and P . mendminu and P. stutzeri. A strain was identified to the combined taxon if the two taxa appeared as first
and second choice and the sum of their identification scores was greater than 0.999.
Strains. We identified 621 strains collected between 1965 and 1983. The aim was to include at least 10 strains of
each taxon (if available) and the type strain (if available) was always included. Many of the strains were from
culture collections, or were designated strains.
ZdentiJicationprocedure. Type strains, strains from culture collections and designated strains were taken as
independently identified. The remaining strains used to evaluate the matrix were identified only by the computer
method using earlier versions of the matrix, although subject to a final assessment. All strains were then assigned
to taxa in the matrix. In most cases a selection of the same strains used to compile the matrix were used to
evaluate it.
After the present version of the matrix was generated, all strain data were resubmitted for computer
identification. This computer decision was compared with the original strain designation. If the computer
identification on the present matrix was the same as the original designation then the computer identificationwas
considered to be correct; if not, then the computer identification was considered to be incorrect. Strains whose
identification scores did not exceed the threshold identificationlevel were considered as unidentified. Any strain
which could not be identified on earlier versions of the matrix, even after review by us, was excluded from this
study. Such intermediate or highly unusual strains which fail to identify are discussed by Lapage et al. (1973).
Strains of taxa not included in the matrix were not evaluated.
Statistical evaluation. The quality of the matrix was tested by two statistical programs. Program OVERMAT
(Sneath, 1980b)measures the overlap between pairs of taxa in the matrix. For each pair, the extent and statistical
significanceof the overlap were determined. Groups which show unacceptably large overlap with others can thus
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Wed, 12 Jul 2017 16:26:54
Probabilistic identification of non-fermenters
1833
be found. Program MOSTTYP (Sneath, 1980a) calculates the best identification score possible for a theoretically
most typical member of the taxon (hypothetical median organism, or HMO). These two programs test the
homogeneity of the taxa on the tests in the matrix.
RESULTS A N D DISCUSSION
Identification of strains
The computer identificationof the various taxa on the present matrix and a comparison with
the respective figures obtained on the matrix of Bascomb et al. (1973) for those taxa included in
both matrices is shown in Table 2. The totals show that 914% of the strains were correctly
identified on the present matrix compared with 82.9% on the matrix of Bascomb et al. (1973).
On the present matrix, all the strains in 49 of the 66 taxa were identified, and 90%or more of the
strains were correctly identified in a further six taxa.
Separate taxa
Hereunder, names given in inverted commas are names that are not included in the Approved
Lists of Bacterial Names (Skerman et al., 1980) nor in their supplements.
Achromobacter. Seven taxa were included - the six Achromobacter groups A to F defined by
Holmes & Dawson (1983) and the named species A. xylosoxidans (see Holmes et al., 1977b). All
strains in Achromobacter groups A to F were of clinical origin and all strains were correctly
identified.Achromobacter species biotypes 1 and 2 of Tatum et al. (1974) were reported as being
separableon the basis of reactions in maltose, mannitol and sucrose, but we found that reference
strains of both these taxa gave positive results in the present study. Both groups have been
included in Achromobacter group A. Of the 14 strains of A . xylosoxidans, 12 were correctly
identified. A. xylosoxidans had the highest identificationscore for the remaining two strains, but
in both, the score for Alcaligenes denitrificans prevented identification at the level of 0.999. This
is not surprising, as these two taxa are so closely related that it has been proposed that they be
recognized as sub-species of the same species, i.e. as Alcaligenes denitrificans subspecies
denitrificans and Alcaligenes denitrificans subspecies xylosoxydans (Kersters & De Ley, 1984).
Acinetobacter. Two species were included - Acinetobacter calcoaceticus and A. lwofli. All
strains of both species were correctly identified.
Agrobacterium. Four taxa were included. The taxa were as defined by Holmes & Roberts
(1981) from a numerical taxonomic study, which was deemed necessary because plant
pathologists had previously classified these bacteria primarily .according to their pathogenic
effect on plants rather than according to a wide range of phenotypic characters. Although
primarily plant pathogens, strains of A. tumefaciens (synonym A . radwbacter) have been
recovered in clinical specimens referred to us, and strains closely resembling the ‘Agrobacterium
yellow group’ have been isolated from peritoneal dialysis fluid (Swann et al., 1985). Strainsof the
latter taxon probably should be transferred from Agrobacterium when a new genus has been
created to accommodate them; they are more closely related to species such as Pseudomonas
paucimobilis. All strains of A. rhizogenes, A. rubi and of the ‘Agrobacteriumyellow group’ were
correctly identified. Of the 12 strains of A. tumefaciens, 11 were correctly identified. A.
tumefaciens had the highest identificationscore for the remaining strain, but the score for A. rubi
prevented identification at the level of 0.999.
Alcaligenes. Two species were included - A. denitrificuns and A . faecalis. They showed close
similarity in phenotypic tests and only nitrate reduction in our matrix provided a 0.99/0-01
separation. These two taxa are therefore allowed to identify together as a composite group if the
sum of the scores for the two taxa exceeds the identificationlevel of 0.999; nevertheless, they are
distinct species and should perhaps be placed in different genera (see Holmes & Dawson, 1983).
‘A. odorans’ was not included in the matrix because it is now recognized as a synonym of A.
faecalis (see Holmes & Dawson, 1983). Four of the five strains of A. denitrificans were identified
correctly (though two of these as a composite group with A. faecalis). For the remaining strain,
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Wed, 12 Jul 2017 16:26:54
1834
B . HOLMES, C . A. PINNING A N D C . A . DAWSON
-
Table 2. Identijcation of non-fermentative bacteria
Final version of matrix
Taxon
no.
*l.
*2.
3.
4.
5.
6.
*7.
*8.
*9.
10.
11.
12.
13.
*14.
*15.
*16.
*17.
*18.
*19.
*20.
*21.
*22.
*23.
*24.
*25.
*26.
*27.
*28.
*29.
*30.
*31.
*32.
*33.
*34.
*35.
*36.
*37.
38.
*39.
*40.
*41.
42.
*43.
*44.
*45.
*46.
*47.
*48.
Taxon
Achromobacter Group A
Achromobacter Group B
Achromobacter Group C
Achromobacter Group D
Achromobacter Group E
Achromobacter Group F
Achromobacter xylosoxidans
Acinetobacter calcoaceticus
Acinetobacter lwofii
Agrobacterium rhizogenes
Agrobacterium rubi
Agrobacterium tumefaciens
Agrobacterium yellow group
Alcaligenes denitriJicans
Alcaligenes faecalis
Alteromonas putrefaciens
Bordetella bronchiseptica
Bordetella parapertussis
BranhamellalM. nonliquefacienslNeisseria
Brucella spp.
CDC Group IIf
CDC Group IIj
CDC Group IVe
CDC Group Ve, type 1
CDC Group Ve, type 2
Eikenella corrodens
Flavobacterium breve
F. meningosept icum
F. mult ivorum
F. odoratum
Flavobacterium species Gp IIb
F. spiritivorum
F. thalpophilum
Janthinobacterium lividum
Kingella denitrijicans
K . indologenes
K . kingae
Moraxella anatipestifer
M.osloensis
M . phenylpyruvica
Moraxella proteolytic group
M . saccharolytica
M . urethralis
Pseudomonas a idovorans
P . aerugimsa
P . alcaligenes
P . cepacia
P . diminuta
No. of
strains
25
16
2
3
2
1
14
11
17
10
7
12
7
5
11
11
10
10
22
11
10
10
4
9
11
10
14
15
10
10
11
10
7
10
3
2
4
1
10
10
12
1
11
11
13
5
12
11
Strains correctly
identified
by computer
Matrix of
Bascomb et al. (1973)
w
Strains correctly
identified
by computer
*
%
No.
25
16
2
3
2
1
12
11
17
10
7
11
7
4
11
11
10
10
4
8
9
10
4
9
11
10
9
13
9
10
9
10
7
10
3
2
4
1
10
10
3
1
11
11
13
4
12
11
100
100
100
100
100
100
85.7
100
100
100
100
92.5
100
80
100
100
100
100
18.2
72.7
90
100
100
100
100
100
64-3
86.7
90
100
81.8
100
100
100
100
100
100
100
’’}
25
100
100
100
137
-
100
39
80
100
100
-$
-
* Taxa recommended for retention in a reduced matrix.
?The Moraxella spp. of the matrix of Bascomb et al. (1973) does not correspond exactly to the range of
Moraxella species included in the final version of our matrix.
2 In the matrix of Bascomb et al. (1973, strains of these taxa were included in ‘Cornamonaspercolans’,a taxon no
longer recognized.
0 These figures are different from those published for the matrix of Bascomb et al. (1973) to take account of taxa
which were not retained in the final version of our matrix.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Wed, 12 Jul 2017 16:26:54
Probabilistic identijication of non-fermenters
1835
Table 2 (continued)
Matrix of
Bascomb et al. (1973)
Final version of matrix
r
Taxon
no.
*49.
50.
51.
*52.
*53.
*54.
*55.
*56.
*57.
*58.
*59.
*60.
61.
*62.
*63.
64.
65.
66.
Taxon
P.Juorescens
P . fiagi
P. lemoignei
P.mallei
P . maltophilia
P. mendocina
P . paucimobilis
P. pickettii
P.pseudoalcaligenes
P . pseudomallei
P. putida
P . stutzeri
P. taetrolens
P . testosteroni
P . uesicularis
Rhizobium meliloti
‘Xanthomonas hyacinthi’
Xanthomonas spp. not hyacinthi
Total
1
Strains correctly
identified
by computer
N0.0f\ - A - t
strains
No.
11
10
1
9
10
10
10
11
11
9
1
9
8
10
10
10
10
10
11
9
10
10
11
3
8
10
10
10
10
62 1
568
10
10
10
12
3
10
10
7
Strains correctly
identified
by computer
%
100
90
100
100
80
100
100
100
90
100
100
91.7
100
80
100
100
100
100
91.5
* Taxa recommended for retention in a reduced matrix.
TThe Moraxella spp. of the matrix of Bascomb et al. (1973) does not correspond exactly to the range of
Moraxella species included in the final version of our matrix.
# In the matrix of Bascombet al. (1973), strainsof these taxa were included in ‘Comamnaspercolans’, a taxon no
longer recognized.
0 These figures are different from those publishedfor the matrix of Bascombet al. (1973) to take account of taxa
which were not retained in the final version of our matrix.
A. denitrijicans had the highest identification score and A. faecalis the second highest score, but
even the combined scores did not exceed the threshold identificationlevel of 0.999. All strains of
A. faecalis were correctly identified, three as a composite group with A. denitrijicans. Other
species of Alcaligenes appearing on the Approved Lists of Bacterial Names (Skerman et al.,
1980) are marine species or hydrogen-utilizing species; these are not included in the matrix.
Alteromonas. One species, A. putrefaciens, was included and all 11 strains were correctly
identified.
Bordetella. Two species were included - B. bronchiseptica and B.parapertussis (B.pertussis was
excluded as strains of this species do not grow on nutrient agar). All 10 strains of B. bronchiseptica
and all 10 strains of B. parapertussis were correctly identified.
Branhamella/Moraxella/Neisseria.
Strains belonging to these genera were largely unreactive in
the tests included in the matrix (Table 1). It is therefore.difficult to separate the species from
each other, even when they are in different genera. A cluster analysis has not been done, to
determine how to group these strains into taxa in the matrix, so a purely intuitive approach was
adopted. Certain Moraxella species had a few characteristics which permitted their easy
differentiationfrom other species in the genus. These taxa, Moraxella anatipestifr, M . osloensis,
M . phenylpyruvica,M . saccharolytica and M . urethralis,were entered separately in the matrix and
all strains of each species were correctly identified. The largely unreactive strains producing
gelatinase, Moraxella bovis, M . equi, M . lacunata and M . liquefaciens, were included in a single
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Wed, 12 Jul 2017 16:26:54
1836
B. HOLMES, C. A . PINNING AND C. A . DAWSON
taxon, designated the Moraxella proteolytic group. However, only three of the 12 strains of this
taxon were correctly identified. Moraxella proteolytic group had the highest identification score
for seven of the remaining strains but identification at the level of 0.999 was prevented by the
scores for BranhamellalM. nonliquefacieslNeisseria in five strains and by the scores for CDC
Group I1j in two strains. For the two remaining strains, BranharnellalM. nonliquefaciens/
Neisseria had the highest identification score.
All the species included in the Moraxella proteolytic group were represented amongst the
strains used to evaluate this taxon in the matrix. The three strains that identified correctly to the
taxon included one of M. equi and one of M. liquefaciens (the remaining strain was a field
isolate). The particularly unreactive species of this group were included in a single taxon,
Branhamella/M. nonliquefaciens/Neisseria. In this taxon were placed the following species :
Branhamella catarrhalis, Moraxella nonliquefaciens, Neisseria animalis, N. canis, N . caviae, N .
cinerea, N. cuniculi, N. elongata, N.fiavescens, N . lactamica, N.meningitidis and N . ovis. For the
following species, which might have been included in this taxon, reference strains were
unavailable at the time the matrix was compiled : Moraxella atlantae, Neisseriaflava, N . perflava,
N. sicca and N . subjlava. Reference cultures of Neisseria denitrificans, N. mucosa and ‘N.
pharyngis’ were found to give a fermentative reaction in the O/Ftest, and so were excluded from
the matrix for non-fermentative organisms; ‘N.pharyngis’ and N . mucosa both produce acid
from glucose, maltose and sucrose in peptone-containing media. N . gonorrhoeae was not
included because the strains do not grow on nutrient agar. The majority of culture collection
strains of N. meningitidis grow on nutrient agar, so N . meningitidis was included in this taxon.
There was a poor identification rate for this taxon with only four of the 22 strains correctly
identified. Branhamella/M. nonZiquefacienslNeisseria had the highest identification score for 15
of the 18 remaining strains but identification at the level of 0.999 was prevented by the score for
Brucella in one strain, for Kingella denitrificans in another, for K. kingae in eight strains, for
Moraxella osloensis in one strain and for the Moraxella proteolytic group in four strains. For the
three remaining strains, the Moraxella proteolytic group had the highest identification score;
for two of these strains, Branhamella/M. iwnliquefaciem/Neisseria had the second highest
identification score, whilst for the remaining strain Moraxella phenylpyruvica had the second
highest identification score.
All the species included in the Branhamella/M. nonliquefaciens/Neisseria group were
represented amongst the strains used to evaluate this taxon in the matrix. The four strains that
identified correctly to the taxon were strains of B. catarrhalis, N. cuniculi, N . elongata and N .
lactamica.
Brucella. This taxon comprised :B. abortus, B, canis, B. melitemis, B. neotomae and B. suis,but
did not include B. bovis, the strains of which failed to grow on nutrient agar. Eight of the 11
strains of Brucella species identified correctly. Bmcella species had the highest identification
score for the remaining three strains but in each case the score for
-_
Branhamella/M.
nonliquefaciens/Neisseria prevented identification at the level of 0.999. Each of the three latter
strains belonged to a different Brucella species: B. abortus, B. neotomae and B. suis.
All the species included in Brucella spp. were represented amongst the strains used to evaluate
this taxon in the matrix. The eight strains that identified correctly to the taxon included strains
of B. abortus (one strain), B. canis (one strain), B. melitensis (three strains) and B. suis (two
strains).
._
CDC Group 11’ This unnamed taxon has been described by Tatum et al. (1974). Strains are
most commonly associated with the female urino-genitary tract. Of 10 strains, nine were
correctly identified. CDC Group IIf had the highest identification score for the remaining strain
but the score for the Moraxella proteolytic group prevented identification at the level of 0.999.
CDC Group IZj. This unnamed taxon has also been described by Tatum et al. (1974). This
organism is commonly isolated from the oral and nasal fluids of healthy dogs (Bailie et al., 1978),
and is frequently recovered from dog-bite wounds. All 10 strains of the taxon were correctly
identified.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Wed, 12 Jul 2017 16:26:54
Probabilistic ident$ca tion of non-fermenters
1837
CDC Group W e ,This unnamed group shares some similarity in biochemical tests to species of
Alcaligenes and most strains have been recovered from urine (Tatum et al., 1974). All four strains
of this taxon were correctly identified.
CDC Group Ve, types I and 2. These yellow-pigmented,oxidase-negativeorganismshave been
recovered from a variety of anatomical sources in humans, but are most frequently associated
with wounds and abscesses (Tatum et al., 1974). Although treated together, the two types are
quite separate species, CDC Group Ve type 1 strains having multitrichous polar flagella and a
G C content of 56-8mol % and Group Ve type 2 strains having a single polar flagellum and a
G C content of 68.9 mol % (Gilardi et al., 1975). All nine strains of the former taxon and all 11
strains of the latter were correctly identified.
+
+
Eikenella. All 10 strains were correctly identified to E. corrodens.
Flavobacterium. Six named species were included together with the unnamed Flavobacterium
species Group IIb (Tatum et al., 1974). Except for the most recently described species, F.
thalpophilum, all the remaining taxa have been described by Holmes et al. (1984~).The species
F. odoratum, F. multivorum, F. spiritivorum and F. thalpophilum (see Holmes et al., 1977c, 1979b,
1981,1982,1983)are phenotypically homogeneous taxa which can be separated from each other
on several phenotypic characters. Not surprisingly, therefore, all strains belonging to these four
species were correctly identified, with the exception of one strain of F. multivorum, where
although the latter taxon achieved the highest identification score, the score for F. thalpophilum
prevented identification at the level of 0.999. The remaining Flawbacterium taxa, however,
show varying degrees of heterogeneity in both phenotypic characters and DNA characteristics
and there are few constant phenotypic characters to differentiate them. These taxa had low
identification rates, despite the allowance of identification to a composite group between F.
meningosepticum and Flatlobacteriumspecies Group IIb. The F. breve1F. meningosepticumlFlatlobacterium species Group IIb complex is known to present taxonomic problems (Holmes, 1983).
Only nine of the 14strains included in F. breve (see Holmes et al., 1978)were correctly identified.
In four of the five remaining strains, F. breve had the highest identification score, but
identificationat the level of 0.999 was prevented by the scores for CDC Group IIf in one strain,
F. meningosepticum in another, and Flavobacterium species Group IIb in the other two strains.
For the final strain Flavobacterium species Group IIb had the highest identificationscore and F.
breve the second highest identification score. Of the 15 strains of F. meningosepticum, 13 were
correctly identified, one as a composite group with Flavobacterium species Group IIb. For one of
the two remaining strains, F. meningosepticurn had the highest identification score but
identification at the level of 0.999 was prevented by the score for F. breue. The remaining strain
identified as one of F. breve and, therefore, either represents an example of misidentificationon
this matrix, or the strain was incorrectly ascribed to F. meningosepticum when assessed by us. Of
the 11 strains of Flavobacterium species Group IIb, 9 were correctly identified, none as a
composite group with F. meningosepticum. As there are no tests in the matrix providing a
0.99/0.01 separation between these two taxa it is interesting that so many strains achieved
identification to their respective taxa at a level of 0.999. For one of the two remaining strains of
Flatlobacterium species Group IIb, this taxon had the highest identification score, but
identification at the level of 0.999 was prevented by the score for F. breve. In the remaining
strain, F. meningosepticum and F. breve achieved respectively the highest and second highest
identification scores, whilst Flauobacterium species Group IIb had the third highest
identification score. This strain is interesting, as in a numerical taxonomic analysis (Holmes,
1983) it clustered with strains of Flavobacterium species Group IIb, but in a more recent
numerical taxonomic study (B. Holmes, unpublished results) it joined a cluster containing
strains of F. breve. The identity of this strain is thus uncertain.
For the unnamed taxon Flavobacterium species Group IIb, which is heterogeneous in both
phenotypic and DNA characteristics, the name F. indologenes has been proposed (Yabuuchi et
al., 1983). For certain strains in Flavobacterium species Group IIb, showing a high level of
DNA-DNA relatedness, the name F. gleum has been proposed (Holmes et al., 1984b). Pending
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Wed, 12 Jul 2017 16:26:54
1838
B. HOLMES, C. A . PINNING AND C . A. DAWSON
further study of this group we recommend retention of the original designation Flavobacterium
species Group IIb.
Other Flavobacterium species appearing on the Approved Lists are not included because they
neither conform to the genus as currently defined (Holmes et al., 1984a), nor appear to occur in
human clinical specimens.
Janthinobacteriurn.The only species, J. lividum, was included; the species was formerly known
as ‘Chrornobacteriumlividum’. Although strains do not grow at 37 “C,they may nevertheless be of
clinical importance as they can grow in refrigerated products such as blood, and could therefore
lead to shock if transfused into a patient. All 10 strains of the species were correctly identified.
Kingella. Three species were included: K. denitrijicans, K . indologenes and K . kingae. All
strains of each species were correctly identified, although one strain of K. denitrijicans and two
strains of K. kingae were identified only to the composite group K. denitrijicans/K. kingae. The
only test in our matrix which provided a 0*99/0.01separation between these two taxa was nitrate
reduction; hence, the allowance of identification to a composite group. Although relatively
inactive in most of the tests in the matrix, the lack of catalase production readily distinguished
Kingella strains from those of Branhamella, Moraxella and Neisseria.
Pseudomonas. Twenty species were included and all strains of each of the following species
were correctly identified : Pseudomonas acidovorans, P . aeruginosa, P. cepaciu, P . diminuta, P .
JEuorescens (none as a composite group with P . putida), P . lemoignei (this species has only been
isolated previously from garden soil), P . mallei, P . mendocina (none as a composite group with P.
stutzeri; P . mendocina has only once been reported from a human clinical specimen), P.
paucimobilis (formerly known as CDC Group IIk, type 1;Holmes et al., 1977a), P . pickettii (this
taxon includes organisms formerly designated as CDC Groups IVd, Va-1, Va-2 and
‘Pseudomonas thomasii’; see King et al., 1979), P . pseudomallei, P. putida (one strain as a
composite group with P.fluorescens), P.taetrolens and P . vesicularis. No strains were received of
P . taetrolens and there were only a few strains of P . diminuta and P. vesicularis from human
clinical specimens. Of the five strains of P . alcaligenes, four were correctly identified. In the
remaining strain, P . alcaligenes had the highest identification score, but the score for P .
testosteroni prevented identification at the level of 0.999. Occasionally strains were received for
identification of P . alcaligenes from human clinical material. Of the 10 strains of P . fragi, nine
were correctly identified. For the remaining strain, P . fragi had the highest identification score
but the score for P . taetrolens prevented identification at the level of 0-999.This species does not
appear to occur in human clinical specimens. Of the 10 strains of P . maltophilia, eight were
correctly identified. In the two remaining strains, P . maltophilia had the highest identification
score, but the scores for Flavobacterium species Group IIb in one strain and Pseudomonas
pseudoalcaligenes in the other prevented identification at the level of 0.999. P . maltophilia,
commonly recovered from blood cultures, is the second most common pseudomonad isolated
from human clinical specimensafter P. aeruginosa (see Holmes et al., 1979a). Swingset al. (1983)
have proposed transferring P . maltophilia to the genus Xanthomonas as Xanthomonas maltophilia.
Of the 10 strains of P . pseudoalcaligenes, 9 were correctly identified. P . pseudoalcaligenes had the
highest identification score in the remaining strain, but the score for P . alcaligenes prevented
identification at the 0.999 level. Of the 12 strains of P . stutzeri 11 were correctly identified, none
as a composite group with P . mendocina. P . stutzeri had the highest identification score in the
remaining strain, but the score for P . pseudoalcaligenes prevented identification at the level of
0-999.There are no tests in the matrix providing a 0-99/0.01 separation between P._Puurescens
and P . putida, and between P . mendocina and P . stutzeri, yet most strains achieve identification
to their respective taxa at a level of 0.999. Of the 10 strains of P . testosteroni, eight were correctly
identified. P. testosteroni had the highest identification score in the two remaining strains, but
the score for P.acidovorans prevented identification at the level of 0.999 in both strains. Other
species of Pseudomonasappearing on the Approved Lists were not included in the matrix as they
are not known to occur in human clinical specimens.
Rhizobium. Only R.meliloti was included. All strains of R. meliloti were identified correctly.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Wed, 12 Jul 2017 16:26:54
Probabilistic ident $cation of non-fermenters
1839
Strains of the genus are closely associated with plants, and only R. meliloti was included because
strains of other Rhizobium species failed to grow on nutrient agar. R. meliloti has not so far,
however, been reported from human clinical specimens.
Xanthomonas. Strains of this genus are primarily plant pathogens. Isolates from human
clinical specimens reported in the literature as belonging to this genus are probably strains of P.
paucimobilis. When compiling the matrix, we examined culture collection strains of
‘Xanthornonas hyacinthi’, X. campestris, ‘X. juglandis’ and ‘X.pelargonii’, but not of other
Xanthomonas species appearing on the Approved Lists. The strains of ‘ X . hyacinthi’ were easily
distinguishable from the other species using the tests in the matrix, but the three latter species
were not easily distinguishable from each other: they are included together in the matrix as a
single taxon, Xanthomonas spp. (not hyacinthi).All strains of both taxa were correctly identified.
Stat ist ical evaluation
If the percentage overlap between two taxa in the matrix is large, an unknown belonging to
one or other taxon will not score highly with either. An overlap of less than 1.0%is desirable
(Sneath, 1980b). All pairs of taxa overlapped by less than this value.
If the matrix is satisfactory the most typical strain of a taxon should give a high identification
score close to 1.000 against its own taxon (Sneath, 1980~).A much lower score should be given
by the second best identity. Of the 66 taxa, four did not achieve a score of 3 0.999 for the HMO;
these were Alcaligenes faecalis (0*998),BranhamellalM. nonliquefacienslNeisseria (0.994), K.
denitrijicans (0.995) and Moraxella proteolytic group (0.998).
G E N E R A L DISCUSSION
The probability matrix of Bascomb et al. (1973) yielded an identification rate of 90-8% for
fermentative bacteria and 82.1 % for non-fermenters. The identification rate for the latter group
of organisms was much lower than that for the fermenters because of the unsuitability of nearly
half the tests in their matrix. The methyl red and Voges-Proskauer tests are invariably negative
with non-fermenting strains and in peptonelwaterlsugar media, saccharolytic non-fermenters
generally give negative results (see also Holmes 8z Hill, 1985). The use of ammonium salt sugars
in our matrix permitted the detection of acid produced by saccharolytic non-fermenters and
discrimination between these taxa improved to 96% (382 out of 400 strains). The other tests
included in the present matrix, but not in that of Bascomb et al. (1973), served to improve
discrimination between the saccharolytic and non-saccharolytic strains. Although fewer tests
were available to discriminate between the non-saccharolytic strains, 186 (84%) of 221 such
strains were successfully identified. This is less satisfactory than for saccharolytic ones, but the
identification rate of 84% is an improvement on the overall identification rate of 82.1%,
achieved for the non-fermenting strains, using the matrix of Bascomb et al. (1973). For nonsaccharolytic taxa such as Acinetobucter lwofii, Alcaligenesfaecalis, Bordetella paraprtussis and
certain Moraxella species the identification rate on our matrix was considerably improved
(Table 2).
Shaw & Latty (1984) described a probability matrix comprising 11 taxa and 18 carbon source
utilization tests, for the identification of Pseudomonas taxa from meat. Their probabilistic
method was similar to our own but with a lower threshold identification level of 0.99. Although
they were identifying meat isolates rather than clinical isolates, most of the taxa were common to
both matrices, and they reported an overall identification rate of 89.7%, which is very similar to
our figure of 913%.
Bascomb et al. (1973) found that it was not clear exactly what factors operate in successful
probabilistic identification. They found that the number of biopatterns given by strains in a
taxon showed no correlation with success in identification in that taxon. However, the number
of 0.99 or 0.01 differences, for a given taxon, from the taxon most similar to it in the matrix was
important, as a successful identification rate cannot be assumed (though it may be obtaiped)
unless two or more 0.99/0.01 test differences exist between the taxa concerned. The only fully
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Wed, 12 Jul 2017 16:26:54
1840
B. HOLMES, C . A . PINNING AND C . A. DAWSON
Table 3. Tests of greatest value for digerentiating the taxa in a reduced matrix
The taxa recommended for retention in a reduced matrix are indicated with asterisks in Table 2. ASS,
ammonium salt/sugar medium; PWS, peptone/water/sugar medium.
1. *Motility at room temperature (growth at room
2.
3.
4.
5.
6.
7.
8.
9.
10.
1I.
12.
13.
14.
15.
16.
17.
18.
temperature)
Acid from ASS glucose
Acid from ASS maltose
Nitrate reduction
*Lipid inclusions after growth on P-hydroxybutyrate (growth on P-hydroxybutyrate, but see
test 36)
Gelatinase production (plate method)
Oxidase production
Tween 80 hydrolysis
Urease production
*Alkaline reaction in Hugh and Leifson O/F test
(fermentative, oxidative or negative reactions
also)
Growth on Simmons' citrate
Casein digestion
Catalase production
Growth at 42 "C
Tween 20 hydrolysis
Arginine dihydrolase production
*Acid from PWS glucose (gas from PWS glucose,
but should be negative)
Acid from ASS trehalose
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31,
32.
33.
34.
35.
36.
37.
Acid from ASS ethanol
Acid from ASS xylose
Nitrite reduction
Lysine decarboxylase production
*Bile stimulation (growth on 10% and 40% bile,
haemolysis)
KCN tolerance
Growth on 4% (w/v) NaCl
Alkali production on Christensen's citrate
Gelatinase production (stab liquefaction)
'\
P-Galactosidase production (ONPG test)
Acid from ASS adonitol
H2S production (triple sugar/iron agar)
Ornithine decarboxylase production
Acid from ASS mannitol
Acid from ASS salicin
Acid from ASS fructose
*Production of brown melanin-like pigment on
tyrosine agar (tyrosine hydrolysis)
*Growth on a-hydroxybutyrate (lipid inclusions
after growth on P-hydroxybutyrate, but see
test 5)
Growth on cetrimide agar
* With these tests, results are automatically available for the additional tests written in parentheses. These
additional tests can thus be included in the reduced matrix.
correlated factor, indicated from the studies of Bascomb et al. (1973), was the number of
differences of the strain from the matrix entries of 0.99 or 0.01 for that taxon. Otherwise it
seemed to them that there was no method apriori in a biological system of determining whether a
probabilistic matrix will operate successfully, and an empirical trial has to be made. We have
done such a trial of our matrix and have also employed two programs designed for the statistical
analysis of probability matrices. These showed that almost all taxa in the matrix were suitably
discrete and achieved a high identification score for their HMO. All showed the desired level of
homogeneity and separation. The results of the statistical evaluation thus corresponded closely
to the results of the empirical trial.
Several of the strains which did not identify on our matrix were ones which varied from the
expected test results in their taxon; such differences are frequently ignored by the microbiologist
when making a decision on the identification of these strains. If the logic of such decisions could
be formulated it might be possible to include it in the identification program. Similarly,
formulation of more precise definitions for taxa would enhance identification and would remove
the weighting of characters, often used in conventional identification. A number of taxa had to
be defined or re-defined before the present matrix could be completed. More work is still
needed, however, on certain taxa of the genera Branhamella, Flavobacterium, Moraxella and
Neisseria and on the non-saccharolytic pseudomonads, where least identification success was
achieved. One possibility would be to discover additional conventional tests to differentiate
these organisms and to incorporate these tests in the matrix. If there were a large number of such
tests, a separate matrix might have to be constructed for these organisms. Alternatively, other
techniques, such as electrophoresis of cellular proteins or gas chromatography, may have to be
used to discriminate adequately between these organisms.
The matrix described here should prove of value in the examination of strains in the reference
laboratory. With the increasing use of micro-computers, some readers may wish to adapt the
matrix for routine laboratory use. This may well involve a reduction in both number of taxa and
number of tests. To aid interested readers, we indicate in Table 2 those taxa of known or
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Wed, 12 Jul 2017 16:26:54
Probabilistic ident$cation of non-fermenters
1841
probable clinical significance that we consider should be retained in any reduced matrix. The set
of tests with the highest differential value for these taxa are indicated in descending order of
differential value in Table 3. This set of 37 tests separates most pairs of taxa in the matrix by at
least two tests. Eleven pairs of taxa are separated by only one test and three by none; these three
pairs are Flavobacterium meningosepticum and Flavobacterium species Group IIb, Pseudomonas
fluorescens and P. putida, and Pseudomonas mendocina and P . stutzeri, all of which pairs are
allowed to achieve composite group identification. The number of results obtainable from these
37 tests is 46 and the additional characters are listed in parentheses in Table 3. However, not all
tests in the reduced matrix need to be done on every isolate; the bile tests, for example, which are
time-consuming to prepare, could be done only on those strains not identified on the initial tests.
We have not constructed such a reduced matrix ourselves so we cannot vouch for its
effectiveness. It is likely that fewer strains would exceed the threshold identification level of
0.999 on the reduced matrix than on the full matrix. This can be overcome by lowering the
threshold level, as has been done by the manufacturers of commercial identification systems
who face similar problems (see, for example, API, 1983). Lowering the identification level,
however, will increase the risk of misidentification (Lapage et al., 1973).
We are deeply indebted to the late W. R. Willcox for developing and operating the various identification and
summary programs upon which this publication is based and to D. M. Shankie-Williams for transferring the
programs to, and operating them on, a CTL 8046 computer. We thank the staff of the NCTC Computer
Identification Laboratory for their technical support, E. Roe for typing the manuscript and R. K. A. Feltham for
carrying out the statistical evaluations.
REFERENCES
API (1983). API20E Analytical Profile Index. La Balme
Les Grottes: API System SA.
BAILIE,W. E., STOWE,E. C. & SCHMITT,A. M. (1978).
Aerobic bacterial flora of oral and nasal fluids of
canines with reference to bacteria associated with
bites. Journal of Clinical Microbiology 7 , 223-23 1,
BASCOMB,
S., LAPAGE,
S. P., CURTIS,M. A. & WILLCOX,
W. R. (1973). Identification of bacteria by computer : identification of reference strains. Journal
of General Microbiology 77,29 1-3 15.
DYBOWSKI,
W. & FRANKLIN,
D. A. (1968). Conditional
probability and the identification of bacteria : a pilot
study. Journal of General Microbiology 54, 21 5-229.
GILARDI,G. L., HIRSCHL,S. & MANDEL,M. (1975).
Characteristics of yellow-pigmented nonfermentative bacilli (Groups VE-I and VE-2) encountered in
clinical bacteriology. Journal of’Clinical Microbiology
1, 384-389.
HOLMES,B. (1983). The taxonomy of the genus Flavobacterium. In Gram Negative Bacteria of Medical
and Public Health Importance : Taxonomy-Identification-Applications. (Proceedings of a symposium,
Lille, May 25-27, 1983.) Edited by H. Leclerc. Les
Editions INSERM 114, 273-294.
HOLMES,B. & DAWSON,C. A. (1983). Numerical
taxonomic studies on Achromobacter isolates from
clinical material. In Gram Negative Bacteria of’
Medical and Public Health Importance : TaxonomyIdentification-Applications. (Proceedings of a symposium, Lille, May 25-27, 1983.) Edited by H. Leclerc.
Les Editions INSERM 114, 331-341.
HOLMES,B. & HILL, L. R. (1985). Computers in
diagnostic bacteriology, including identification. In
Computer-assisted Bacterial Systematics (Society for
General Microbiology Special Publication no. 1 9 ,
pp. 265-287. Edited by M. Goodfellow, D. Jones &
F. G. Priest. London & New York : Academic Press.
HOLMES,B. & ROBERTS,P. (1981). The classification,
identification and nomenclature of agrobacteria.
Incorporating revised descriptions for each of
Agrobacterium tumejaciens (Smith & Townsend)
Conn 1942, Agrobacterium rhizogenes (Riker et a,.)
Conn 1942, and Agrobacterium rubi (Hildebrand)
Starr & Weiss 1943. Journal of Applied Bacteriology
50, 443-467.
OLMES, B., LAPAGE,s. P. & MALNICK,H. (1975).
Strains of Pseudomonas putrejaciens from clinical
material. Journal of Clinical Pathology 28, 149-1 55.
OLMES,B., OWEN,R. J., EVANS,A., MALNICK,H. &
WILLCOX,W. R. ( 1977a). Pseudomonas paucimobilis,
a new species isolated from human clinical specimens, the hospital environment, and other sources.
International Journal of Systematic Bacteriology 27,
133- 146.
HOLMES,B., SNELL,J. J . S. & LAPAGE,S. P. (1977 b).
Strains of Achromobacter xylosoxidans from clinical
material. Journal of Clinical Pathology 30,595-601.
HOLMES,B., SNELL,J. J. S. & LAPAGE,S. P. (1977~).
Revised description, from clinical isolates, of Flavobacterium odoratum Stutzer and Kwaschnina 1929,
and designation of the neotype strain. International
Journal of’ Systematic Bacteriology 27, 330-336.
HOLMES,B., SNELL,J . J. S. & LAPAGE,S. P. (1978).
Revised description, from clinical strains, of Flavobacterium breve (Lustig) Bergey et al. 1923 and
proposal of the neotype strain. International Journal
of Systematic Bacteriology 28, 201 -208.
B. G.
HOLMES,B., LAPAGE,S. P. & EASTERLING,
(1979a). Distribution in clinical material and identification of Pseudomonas maltophilia. Journal of
Clinical Pathology 32, 66-72.
HOLMES,B., SNELL,J. J. S. & LAPAGE,S. P. (1979b).
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Wed, 12 Jul 2017 16:26:54
1842
B. H O L M E S , C. A . P I N N I N G A N D C. A . DAWSON
Flavobacterium odoratum : a species resistant to a
wide range of antimicrobial agents. Journal of
Clinical Pathology 32, 73-77.
HOLMES,B., OWEN,R. J. & WEAVER,R. E. (1981).
Flauobacterium multivorum, a new species isolated
from human clinical specimens and previously
known as Group IIk, biotype 2. International Journal
of Systematic Bacteriology 31, 21-34.
HOLMES,B., OWEN,R. J. & HOLLIS,D. G. (1982).
Flavobacterium spiritivorum, a new species isolated
from human clinical specimens. International Journal
of Systematic Bacteriology 32, 157-165.
HOLMES,B., HOLLIS,D. G., STEIGERWALT,
A. G.,
PICKETT,M. J. & BRENNER,
D. J. (1983). Flavobacterium thalpophilum, a new species recovered from
human clinical material. International Journal of
Systematic Bacteriology 33, 677-682.
HOLMES,
B., OWEN,R.J. & MCMEEKIN,
T. A. (1984~).
Genus Flavobacterium Bergey, Harrison, Breed,
Hammer and Huntoon 1923,97. In Bergey’s Manual
of’ Systematic Bacteriology, vol. 1, pp. 353-361.
Edited by N. R. Krieg. Baltimore: Williams &
Wilkins.
HOLMES,B., OWEN,R. J., STEIGERWALT,
A. G. &
BRENNER,D. J. (I 984b). Flauobacterium gleum, a
new species found in human clinical specimens.
International Journal of’ Systematic Bacteriology 34,
21-25.
UGH, R. & LEIFSON,E. (1953). The taxonomic
significance of fermentative versus oxidative metabolism of carbohydrates by various Gram negative
bacteria. Journal of’ Bacteriology 66, 24-26.
ERSTERS, K . & DE LEY,J. (1984). Genus Alcaligenes
Castellani and Chalmers 1919, 936. In Bergey’s
Manual of Systematic Bacteriology, vol. 1, pp. 361373. Edited by N. R. Krieg. Baltimore: Williams &
Wilkins.
KING, A., HOLMES,
B., PHILLIPS,I. & LAPAGE,S. P.
(1979). A taxonomic study of clinical isolates of
Pseudomonaspickettii, ‘ P . thomasii’ and ‘Group IVd’
bacteria. Journal of’ General Microbiology 114, 137147.
LAPAGE,S. P. & BASCOMB,
S. (1968). Use of selenite
reduction in bacterial classification. Journal of
Applied Bacteriology 31, 568-580.
LAPAGE,S. P., BASCOMB,S., WILLCOX,W. R. &
CURTIS,M. A. (1970). Computer identification of
bacteria. In Automation, Mechanization and Data
Handling in Microbiology (Society for Applied Bacteriology Technical Series no. 4), pp. 1-22. Edited
by A. Baillie & R. J. Gilbert. London: Academic
Press.
S., WILLCOX,W. R. &
LAPAGE,S. P., BASCOMB,
CURTIS,M. A. (1973). Identification of bacteria by
computer : general aspects and perspectives. Journal
of General Microbiology 11, 273-290.
SHAW,B. G. & LATTY,J. B. (1984). A study of the
relative incidence of different Pseudomonas groups
on meat using a computer-assisted identification
technique employing only carbon source tests.
Journal of Applied Bacteriology 57, 59-67.
SKERMAN,
V. B. D., MCGOWAN,
V. & SNEATH,
P. H. A.
(1980). Approved lists of bacterial names. International Journal of’Systematic Bacteriology 30,225-420.
SNEATH,
P. H. A. (1980~).Basic program for determining the best identification scores possible from the
most typical examples when compared with an
identification matrix of percent positive characters.
Computers and Geosciences 6, 27-34.
SNEATH,
P. H. A. (19806). Basic program for determining overlap between groups in an identification
matrix of percent positive characters. Computers and
Geosciences 6 , 267-278.
SWANN,R. A., FOULKES,
S. J., HOLMES,
B., YOUNG,
J. B., MITCHELL,
R. G. & REEDERS,S. T. (1985).
’Agrobacterium yellow group’ and Pseudomonas
paucimobilis causing peritonitis in patients receiving
continuous ambulatory peritoneal dialysis. Journal of
Clinical Pathology 38, 1293- 1299.
SWINGS,J., DE Vos, P., VANDENMOOTER,M.& DE
LEY,J. (1983). Transfer of Pseudomonas maltophilia
Hugh 1981 to the genus Xanthomonas as Xanthomonas maltophilia (Hugh 1981) comb. nov. International Journal of’SystematicBacteriology 33,409-4 l 3.
TATUM,
W. H., EWING,W. H. & WEAVER,
R.E. (1974).
Miscellaneous Gram-negative bacteria. In Manual
of Clinical Microbiology, 2nd edn, pp. 270-294.
Edited by E. H. Lennette, E. H. Spaulding & J. P.
Truant. Washington, DC : American Society for
Microbiology.
WILLCOX,W. R., LAPAGE,S. P., BASCOMB,S. &
CURTIS,M. A. (1973). Identification of bacteria by
computer : theory and programming. Journal of
General Microbiology 77, 317-330.
S. P.& HOLMES,
B. (1980). A
WILLCOX,
W. R.,LAPAGE,
review of numerical methods in bacterial identification. Antonie van Leeuwenhoek 46, 233-299.
E., KANEKO,
T., YANO,I., Moss,C. W. &
YABUUCHI,
MIYOSHI,N. (1983). Sphingobacterium gen. nov.,
Sphingobacterium spiritivorum comb. nov., Sphingobacterium multivorum comb. nov., Sphingobacterium
mizutae sp. nov., and Flavobacterium indologenes sp.
nov. : glucose-nonfermenting Gram-negative rods in
CDC Groups IIK-2 and IIb. International Journal of’
Systematic Bacteriology 33, 580-598.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Wed, 12 Jul 2017 16:26:54