Kaledin and Verbitsky have shown that symplectic varities have a remarkably nice deformation theory *as symplectic varieties.*

Let $X$ be a symplectic variety (a smooth quasi-projective variety over $\mathbb{C}$ equipped with a nondegenerate closed algebraic 2-form $\Omega$); then there is a universal formal deformation $\tilde X$ over $H^2(X;\mathbb{C})$ completed at the class $[\Omega]$.  

In many examples, it seems that this deformation is generically affine; that is, its fiber at the generic point is an affine scheme.  For example, $T^*G/B$ deforms to a generic coadjoint orbit in $\mathfrak{g}^*$, and if $X$ is a hyperkähler quotient, like a quiver variety, then the deformation comes from varying the complex moment map, and the deformation is also generically affine (since a GIT quotient by a free action of a reductive group is always affine).

Of course, I don't think this always happens; the product of two elliptic curves is symplectic, but doing this deformation should just change the $j$-functions of the curves simultaneously.  So, clearly one needs some kind of extra condition.  I've opted for "resolution of singularities of its affinization."

So my question is:
>Let $X$ be a symplectic variety which is a resolution of singularities of its affinization.  Is $\tilde X$ generically affine?

**EDIT:** As was pointed out in Misha Verbitksy's answer below, the deformation is not canonically algebraic.  I believe though that if you assume that $X$ has a $\mathbb{C}^*$-action which is dilating (Definition 1.7 of  [this paper](http://arxiv.org/pdf/math/0608143.pdf)), then $\tilde X$ will also have a $\mathbb{C}^*$-action which acts by dilation on $H^2(X;\mathbb{C})$ and there will be a unique algebraic structure for which weight vectors of the $\mathbb{C}^*$-action are algebraic functions.  It is this algebraic structure I want.

**EDIT:** Since it's buried a little bit in comments, let me just put here that it's true and proven by Kaledin in [this paper](http://arxiv.org/abs/math/0312134).