In order to put human DNA inside a bacteria in order to have it create Insulin, from what type of cell would you need to take the gene for insulin? I thought it should be from any somatic cell, since the DNA is identical, but I'm told it should be only from the pancreas, why is that?
In short, because the easiest way to get the protein coding sequence of the gene is to create cDNA based on the mRNA, and insulin mRNA is only expressed in pancreatic cells.
Insulin gene consists of two exons. That means, amplifying the genome will not give us a coding sequence -- two pieces of that sequence will be interrupted by a large, non-coding intron.
On the figure above, you see the translated (i.e., coding a protein) sequences represented as opaque boxes. They are linked by a slightly curving thin line -- this is a fragment of sequence that will get excised during mRNA maturation.
To get the coding sequence in a straightforward manner, one can create cDNA from mRNA using a reverse transcriptase. Processed mRNA does not contain the intron. And clearly, pancreas cells express the insulin gene mRNA at very high levels, enabling such a procedure, whereas it is practically not expressed by other tissues.
EDIT: Since I see these comments upvoted, I need to debunk them here:
There are introns and splicing in both, bacteria and archea. Both type I and type II introns are present in bacteria. See this review on bacterial type II introns.
Bacteria also have post-translational modifications, including glycosylation. Once thought to be rare, there is an increasing body of evidence that it plays a much more important role than previously thought. See a paper from a decade ago here.