Is it possible to have a SNP that is not found is be non-randomly associating with any other SNP?
Linkage disequilibrim is a result of two SNPs both appearing after a cross (e.g. in offspring). As such even at great distances two SNPs could both appear in offspring 50% of the time. Linkage refers to a non-random odds that two SNPs (or any other marker) will both appear together in offspring. (Equilibrium being essentially random linkage).
By this definition, SNPs on different chromosomes should have no linkage. Typically linkage reflects that there is a chance of a crossover between the two SNPs that would cause them to not co-migrate. The closer together in general the linkage is very strong - a crossover appearing between two mutations that are just a few bases apart is nearly impossible.
There can be a number of reasons for Linkage Disequilibriunm (LD). Here is a non-exhaustive (non-referenced also) list.
- Random LD happens
- A recent selective sweep
- Population structure
- Genetic drift can also create negative LD (see Barton and Otto, 2005)
Of course, in all cases, reduced recombination rate will help to create an LD and to retain it for a longer period of time.