During the generation of gametes (sperm, eggs), chromosomes can cross-over - this swaps paternal and maternally-derived genetic material. So none of the descendant's chromosomes would be a direct copy of Einstein's, and furthermore, each offspring receives half of its genetic complement from each parent. This means that if you pool the DNA data from small number of offspring you may not obtain all the elements of one parent's DNA or be able to assign that data back to a specific 'original' chromosome. (On top of this you would have to determine which sequences in the offspring came from the other parent).
If I was cloning Einstein, I'd use his DNA. Is his comb in a museum ? A tooth? A letter he licked? A large chunk of Einstein still exists (formalin fixed brain) and lab protocols can recover DNA from formalin fixed tissues. It wouldn't be easy, but in theory, you could: i) derive its full sequence and ii) synthesize artificial chromosomes with suitable epigenetic modifications, and iii) inject those into enucleated eggs for human cloning or iv) transfect them into enucleated cells from a cell line, make iPSCs and then fuse those with ~8-cell fertilized embryos.
(Spoiler - not technically do-able today and almost certainly illegal in most states/countries).
For a related movie treatment, the Boys from Brazil tells a good story - warning Nazi content.