I have started to study the AdS/CFT correspondence by reading Minahan's introductory review (arXiv:1012.3983). The first equation of that review is the expression for the leading contribution to the $\beta$-function:
$$
\beta(g)=-{g^3\over16\pi^2}\left({11\over3}N-{1\over6}\sum_iC_i-{1\over3}\sum_j\tilde C_j\right),
$$
where $C_i$ are quadratic Casimirs due to bosons and $\tidle C_j$ are those due to fermions. The author cites Gross and Wilczek (1973) as a source of the formula. But there it looks a bit differently:
$$
\beta(g)=-{g^3\over16\pi^2}\left({11\over3}N-{4\over3}\sum T_j\right),
$$
where $T_j={d(R_j)\over d(G)}\tilde C_j$, $d(R)$ and $d(G)$ are dimensions of the representation and of the group.
Forget about the boson contribution (Gross and Wilczek were not interested in it). But the fermion contribution contains there the factor $-4/3\times{d(R_j)\over d(G)}$ instead of $-1/3$ in Minahan's article. The factor 2 is related to the fact that Gross and Wilczek considered Dirac fermions, while Minahan considers the Weyl fermion. Ok. But we are left with the extra $2{d(R_j)\over d(G)}$ factor, which I am unable to cancel.
Could anybody explain me this factor? May it be related somehow to supersymmetry?