|
Selectivity of
QTLs.
Before mapping QTLs, we need to consider the
issue of specificity of gene action. The brain weight data in the MBL
were not corrected for the significant differences in the mean body
weights of these strains. Because brain weight may simply be a function
of body size, there is a risk of mapping body weight QTLs instead of
QTLs that have specific effects on brain weight or cell number. To
ensure that brain weight and not some other factor was mapped, variation
in brain weight predictable from variation in body weight must be
factored out. A crude way of doing this is to take the ratio of brain to
body weight as the variable, but a computationally and conceptually
better approach uses regression analysis to remove predictable variance
associated with body size . Williams and colleagues used multiple
regression to remove variance in cell number that was actually
associated with total brain weight. This same logic applies in mapping
QTLs that affect particular CNS cell populations.
The lesson to be taken from this discussion is that whatever types of
QTLs we are trying to map, we need to carefully consider the
higher-order structures and make sure that we have taken variation in
these structures into account.
|
