Dear Raja,

Perhaps the easiest thing to start with is to test for the presence of a global molecular clock in your data. This test is implemented as a standard analysis in HyPhy (under the Molecular Clock rubrik), and it will tell you whether or not a molecular clock model can be rejected for the entire tree.

However, this test will not tell you which branches are responsible for the deviations. You could try the local molecular clock analysis (also implemented in HyPhy) to see which complete subtrees are clocklike and which are not to help pinpoint the source of the deviation.

Is it important for you to know the actual branches, or just that the clock has been violated?

Best,
Sergei