Analysis
proceeded
by
constructing
scatter
plots
allow- ing
serogroups
to
be
compared
on
the
basis
of distribution
of VGs
and
ARI
scores.
ARI
scores
for
the
major
serogroups were
compared,
using
one-way
analysis
of

variance. Univariate
strength-of-association
between
occurrence
of individual
ARGs,
individual

VGs
and

serogroups

were assessed
using
the
relative
risk
statistic
as
follows.
For
all possible
pairings
of
each
resistance
gene,
Ri
(where
i
=
1–28), and
each
VG,
Vk

(where
k
=
1–26),
the
relative
risk
and corresponding
95%
confidence
limits
were
calculated.
In
this context,
the
relative
risk
measures
the
proportion
of
Ri

+ve isolates
that
have
virulence
factor
Vk

divided
by

the proportion
of
Ri


ve
isolates
that
have
virulence
factor
Vk. Relative
risk
measures
how
many
more
times
likely
it
is
for an
isolate
with
resistance
factor
‘‘i’’
to
also
carry
virulence factor
‘‘k’’
compared
to
an
isolate
that
is
not
carrying resistance
factor
‘‘i’’.
Because
of
the
large
number
of
possible combinations
of
Ri

and
Vk

(728
in
all)
the
process
of calculating
all
possible
relative
risks
and
confidence intervals
was
automated
in
the
analysis
package.
Compar- isons
between
the
two
major
ETEC
serogroups
were
made using
relative
risk,

for

example,

possession
of
specific resistance
and
virulence
traits.
Analysis
and
calculation
of the
ARI
was
performed
in
STATA
version
10.1
(STATA Corporation,
College
Station,
Texas,
USA).
VG是毒力基因、AG是耐药基因！