Large samples of over 200 genes are not helpful
since if all of them are correlated with a particular
class, it is unlikely that they all represent
different biological mechanisms and hence are
unlikely to add information not already provided
by others
Though excellent research work has been carried out
in last three decades, there are no general approach
for identifying new cancer classes (class discovery)
or for assigning tumors to known classes
(class prediction)
Conclusion
14 years ago
No comments:
Post a Comment