A neat way of deciding which higher-dimensional theory to start from is to count the number of spinor components.  In ten dimensions, a spinor has $2^{[d/2]} = 32$ components which are reduced to 16 after imposing the Majorana-Weyl condition (which can only be done where d mod 8 = 2). Therefore, $\mathcal{N}=1$ SYM in ten dimensions has 16 components (real !) 

A Majorana spinor in four dimensions has 4 components. So, suppose you want 
$\mathcal{N}=4$ SYM in four dimensions (which has 16 [four copies of four components) you want to start with $\mathcal{N}=1$ SYM in ten dimensions. 

$\mathcal{N}=1$ SYM in four dimensions similarly has 4 components. So, if I want to construct a supersymmetric theory in two dimensions with four supercharges i.e $\mathcal{N}$ = (2,2) SYM (which has single component after imposing Majorana-Weyl), then one should start from $\mathcal{N}=1$ SYM in four dimensions and not with $\mathcal{N}=4$ SYM.