Our bodies are made of about 250 differentiated cell types.

Heart muscle cells are an example of a differentiated cell type.
Red blood cells are another example.
The red-sensitive cone cells of the retina could be considered to be a separate differentiated cell type than the blue-sensitive cone cells and the yellow-sensitive cone cells.

But all retinal ganglion cells would be considered the same cell type, despite the adhesive differences between them that cause their axons to connect preferentially to different locations in the mid-brain. (That is an arbitrary decision, that we may eventually want to change.)

Each differentiated cell type makes a specific subset of proteins.

Once a cell has differentiated into a cell type, it strongly resists switching to being any other cell type. (Differentiation is self-perpetuating)

Anyone would assume that the mechanism of this self-perpetuation would be known, but it isn't, although there are several theories, which include methylation of cytosine in DNA and acetylation of histones. Another possibility is transcription factor genes that stimulate their own transcription, and also stimulate transcription of all the special sub-set of genes expressed by cells of a given type.

(This field might be helped if we didn't have to say "different differentiated" so much.
Awkward vocabulary and other awkward notations (e.g. Java) weaken reasoning powers.)

No cell can differentiate into two or more cell types at the same time.
Some cancer cells are partial exceptions to this rule.
Very little research is done on this subject.

Fusing 2 cells of different differentiated cell types results in them ceasing transcription of both subsets of special "luxury" genes.
I know two exceptions to this rule, in the sense of one nucleus switching to transcription of the same subset of genes.

* If you fuse bird red blood cells with any cell type (because rbc nuclei are inactivated)
* When (very multinucleate) skeletal muscle cells fuse with cells of any other type

With very few exceptions, all differentiated cells contain exactly all the same DNA sequences.

Switching from being one cell type to being another cell type is rare and called "metaplasia"

Switching is fairly common in plant cells; cuttings from a hedge can grow into entire plants. "Rooting" of a cutting from a plant stem is a form of metaplasia, but that is a medical term.

If you could take cells from a person's body, grow them in tissue culture, and stimulate them to switch from one differentiated cell type to another, that would be much better than what "embryonic stem cells" are supposed/hoped to be able to do.

* It would allow you to repair or replace damaged organs.
* The replacement cells would not be attacked by the person's immune system.

My impression is that arguments against using human embryos for religious/moral/political reasons somehow prevented carefully asking whether the method would really work (using mouse embryos, for example.)

QUESTION: A) What property do a few hundred genes need to have in order to be transcribed (only?) in cells of each particular differentiated cell type?

POSSIBLE ANSWER: Maybe they all have some particular base sequence in their promoter and/or enhancer regions.
(There is much evidence for this, but still not conclusive)

ANOTHER QUESTION: What molecular mechanism "turns off" all the luxury genes transcribed by all other differentiated cell types.

POSSIBLE ANSWER? I can't think of a mechanism, and have never heard a good suggestion for one.

Notice that some oncogene mutations (especially translocations) only cause cancer in certain differentiated cell types; Therefore, dedifferentiation might convert such cells to non-cancerous behavior.
(Many or most lymphomas (and leukemias?) result from translocations of oncogenes to the location of the genes for antibodies.

Some More QUESTIONS:

1) Is there a relation between germ layer origin and DNA sequences in promotor regions?
Or similarities in whatever mechanism produces selective transcription of genes?

For example, all differentiated cell types derived from mesoderm might have a particular DNA sequence in the promoter regions of all the genes transcribed in these cell types.

2) The evolutionary origin of a new differentiated cell type occurs how?
Does a new cell type split off from a previous cell type?
Or do two new differentiated cell types split apart, each taking part of the functions, and transcribing a subset of the same genes as a previously-formed cell type?
(Several writers have hypothesized this as the only possibility; and indeed, I can't invent any other.

3) (How) can these mechanisms be deduced / tested using an organism's complete DNA base sequence?

To be more specific, what should one look for (or program your computer to search for) for the purpose of confirming / disproving each theory mentioned above?