<s>
The	O
Z-factor	B-General_Concept
is	O
a	O
measure	O
of	O
statistical	O
effect	B-Application
size	I-Application
.	O
</s>
<s>
The	O
Z-factor	B-General_Concept
is	O
an	O
attempt	O
to	O
quantify	O
the	O
suitability	O
of	O
a	O
particular	O
assay	O
for	O
use	O
in	O
a	O
full-scale	O
,	O
high-throughput	O
screen	O
.	O
</s>
<s>
The	O
Z-factor	B-General_Concept
is	O
defined	O
in	O
terms	O
of	O
four	O
parameters	O
:	O
the	O
means	O
(	O
)	O
and	O
standard	B-General_Concept
deviations	I-General_Concept
(	O
)	O
of	O
both	O
the	O
positive	O
(	O
p	O
)	O
and	O
negative	O
(	O
n	O
)	O
controls	O
(	O
,	O
,	O
and	O
,	O
)	O
.	O
</s>
<s>
Given	O
these	O
values	O
,	O
the	O
Z-factor	B-General_Concept
is	O
defined	O
as	O
:	O
</s>
<s>
The	O
following	O
interpretations	O
for	O
the	O
Z-factor	B-General_Concept
are	O
taken	O
from	O
:	O
</s>
<s>
Z-factor	B-General_Concept
Interpretation1.0Ideal	O
.	O
</s>
<s>
Z-factors	B-General_Concept
can	O
never	O
exceed	O
1.between	O
0.5	O
and	O
1.0An	O
excellent	O
assay	O
.	O
</s>
<s>
Note	O
that	O
if	O
,	O
0.5	O
is	O
equivalent	O
to	O
a	O
separation	O
of	O
12	O
standard	B-General_Concept
deviations	I-General_Concept
between	O
and	O
.between	O
0	O
and	O
0.5A	O
marginal	O
assay.less	O
than	O
0There	O
is	O
too	O
much	O
overlap	O
between	O
the	O
positive	O
and	O
negative	O
controls	O
for	O
the	O
assay	O
to	O
be	O
useful	O
.	O
</s>
<s>
Note	O
that	O
by	O
the	O
standards	O
of	O
many	O
types	O
of	O
experiments	O
,	O
a	O
zero	O
Z-factor	B-General_Concept
would	O
suggest	O
a	O
large	O
effect	B-Application
size	I-Application
,	O
rather	O
than	O
a	O
borderline	O
useless	O
result	O
as	O
suggested	O
above	O
.	O
</s>
<s>
For	O
example	O
,	O
if	O
σp	O
=	O
σn	O
=	O
1	O
,	O
then	O
μp	O
=	O
6	O
and	O
μn	O
=	O
0	O
gives	O
a	O
zero	O
Z-factor	B-General_Concept
.	O
</s>
<s>
The	O
constant	O
factor	O
3	O
in	O
the	O
definition	O
of	O
the	O
Z-factor	B-General_Concept
is	O
motivated	O
by	O
the	O
normal	O
distribution	O
,	O
for	O
which	O
more	O
than	O
99%	O
of	O
values	O
occur	O
within	O
3	O
standard	B-General_Concept
deviations	I-General_Concept
of	O
the	O
mean	O
.	O
</s>
<s>
Another	O
issue	O
is	O
that	O
the	O
usual	O
estimates	O
of	O
the	O
mean	O
and	O
standard	B-General_Concept
deviation	I-General_Concept
are	O
not	O
robust	O
;	O
accordingly	O
many	O
users	O
in	O
the	O
high-throughput	O
screening	O
community	O
prefer	O
the	O
"	O
Robust	O
Z-prime	O
"	O
which	O
substitutes	O
the	O
median	O
for	O
the	O
mean	O
and	O
the	O
median	O
absolute	O
deviation	O
for	O
the	O
standard	B-General_Concept
deviation	I-General_Concept
.	O
</s>
<s>
One	O
estimate	O
of	O
SSMD	B-General_Concept
is	O
robust	O
to	O
outliers	O
.	O
</s>
