Tools
CGAP Data
Purchase CGAP Reagents
Overviews
Related Links
Quick Links:
|
Sequence Odds Ratio and Significance Test
The formula for the sequence odds ratio is:
Seqs A / Total Sequences in A
Seqs B / Total Sequences in B
The results of the DGED are ordered
by the odds ratio, with all cases of "NaN" coming at the head of the list.
"NaN" stands for "not a number" and occurs when the denominator of the
equation is 0, i.e., there are no sequences of a gene in pool B.
To measure the significance of differences in transcript expression, we follow
the method described in
A public database for gene expression in human cancers.
Lal A, Lash AE, Altschul SF, Velculescu V, Zhang L, McLendon RE, Marra MA, Prange C, Morin PJ, Polyak K, Papadopoulos N, Vogelstein B, Kinzler KW, Strausberg RL, Riggins GJ.
Cancer Res. 1999 Nov 1;59(21):5403-7.
Briefly, the method uses Bayesian analysis to compute the probability
that sampling error accounts for the fact that a tag has been
observed to occur more frequently in one library pool than in a second
library pool.
The method uses an assumed prior distribution of tag frequency
and the observed frequencies to
compute the posterior
probability, P, that
the underlying frequency of a tag in one collection of tissues
exceeds, by some factor F, the underlying frequency of the tag
in a second collection of tissues.
The value of P ranges from
0% (very likely that the results are not due to sampling error) to
100% (very likely that the results are
due to sampling error).
Examples of combinations of F and P:
F |
P |
Results |
3 |
.20 |
Moderate probability that the
tag occurs three times as frequently in one pool as in the other |
3 |
.01 |
Strong probability that the
tag occurs three times as frequently in one pool as in the other |
1 |
.20 |
Moderate probability that the
tag occurs more frequently (even only slightly more frequently)
in one pool than
in the other |
1 |
.01 |
Strong probability that the
tag occurs more frequently (even only slightly more frequently)
in one pool than
in the other |
|
|