Well, R1a, I1, I2 whatever, those are all arbitrary things.
Why split at level R1a/R1b? Perhaps we should look higher at R1 or go deeper into R1a subclades.
Then we should also look at similarities of level NO, P, IJ, etc.
Actually the defining mutations of major haplogroups are not all that arbitrary. Most top level haplogroups suffered major bottlenecks during the Last Glacial Maximum. If you look at the phylogenetic tree (either on
Yfull.com or
Isogg.org), you will see that haplogroups are defined by
a lot of mutations, often over 100, while subclades have just a few or even just one. For example (the mutation in brackets is just one of the defining mutations, usually the first identified or the most well known):
E1b1b (M125) : 148 defining mutations
- E1b1b1a (M78) :: 77 defining mutations
G : over 300 defining mutations
-G2a (P15) : 50 defining mutations
I1 : 301 defining mutations
I2 : 64 defining mutations
-I2a1 (M436) 55 defining mutations
-I2a2 (P37) : 28 defining mutations
J1 : 185 defining mutations
J2 : 30 defining mutations
-J2a (M410) : 117 defining mutations
R1a : 102 defining mutations
R1b (M343) : 28 defining mutations
-R1b-V88 : 75 defining mutations
-R1b-P297 : 33 defining mutations
--R1b-M269 : 100 defining mutations
These are of course cumulative and I skipped some intermediary subclades.
Additionally,
as I mentioned before, some haplogroups carry mutations in the coding section of important genes, and therefore are much more likely to cause visible changes in looks or behaviour. This is the case for haplogroups BT, DE, E*, J*, R*, R1a1, R1b-SRY2627, R1b-M222, and T*. I don't think it is a coincidence that these mutations happen to fall right in the node of major historical expansion and define top level haplogroups or very major subclades (R1a1, which means nearly 100% of all R1a people alive today). Even R1b-M222, which is a very young subclade (TMRCA 1900 years before present) now makes up a very considering part of paternal lineages in the British Isles and in the English-speaking world. What exactly is the evolutionary advantage of M222 men, I don't know, but there seems to be one, otherwise this subclade among dozens of L21 subclades would not have become so successful.
The number of mutations between R1b1b (P297) and R1b1b2 (M269) is absolutely stunning. Only about 3500 years elapsed between the two, but M269 acquired 100 new mutations, about 10 per generation ! I cannot think of any other haplogroup that developed such large number of mutations in such a short time. Now is it a coincidence that that very lineage suddenly expanded from a minor Caucasian or Steppe cow herder lineage to a world dominant lineage found on all continents in just a few thousands years? I don't think so. Some of these mutations must have had an effect on the behaviour of these R1b men. We are talking about 100 mutations here, three times more than Paleolithic I* got to become I2a2 over 20,000 years later ! That's not a minor change.
The Y chromosome is the fastest evolving chromosome. Despite the fact that we share 98% of our genome with chimpanzees, our Y chromosomes have already become 30% different. It has been suggested that this is because of the competitive nature of reproduction and that the most advantageous Y chromosomes get selected. However if reproductive fitness was the only role of Y-DNA, we should expect little change over time. If the machinery works properly, why change it? Mutations are actually more likely to damage a fertility tried and tested over millions of years. Surely Y-DNA has other evolutionary implications. One of the biggest differences between the chimpanzees, the gorillas and humans is their respective sexual behaviours. Chimps are extremely promiscuous and use sex as a entertainment and as a bargaining tool, while gorillas are very faithful but live in harems around a dominant male. Humans are somewhere in between, but human behaviour also varies considerably by region and by historical culture. Did Y-DNA influence male behaviour before religions, cultures and laws started regulating sexual behaviour? I think it's an idea worth exploring.