We pretend that the genes we insert are placed at exactly the same spot from one organism to another. That often isn't true, especially when using the earlier insertion mechanisms such as piggyBac and zinc fingers, compared to Crispr/CAS.
While it's a game-changer in gene editing technology, CRISPR-Cas9 isn't error proof.
It has a much higher success rate than the other nuclease technologies when it comes to cutting DNA at the right place. But like the other systems it can also make unintended cuts outside the target gene.
PiggyBac and Zinc Fingers were the first tools that KBLB used to insert genes in silkworms. Note the 'cutting DNA at the right place' comment. If one silkworm has the gene inserted at the correct place and a different silkworm has the gene inserted somewhere else, crossover can create an offspring that does not have the correct genetic sequence.
Note this happens even though the parent is homozygous for the genes we have inserted.
Also note that the genetic editing can impact other genes in the organism, at random.
Over the course of many generations, the locations of genes in non-GMO organisms tend to match up quite precisely in their locations. Our genetic insertions are imperfect and thus we don't have homozygous offspring with homologous gene locations. Crossover can damage the inserted genes (and others) in such organisms. These problems were commonplace in early editing techniques such as piggyBac and zinc fingers. They are more rare with Crispr/CAS.
As usual, KBLB doesn't share enough for us to figure out what is truly going on. Did this problem affect strains created with piggyBac or zinc fingers? Did it affect Crispr/Cas insertions? Calling it 'genetic drift' is a misnomer, as another poster (perhaps Jazz?) indicated.
Here are the possible causes of the situation KBLB is experiencing:
1) Inter-breeding across KBLB genetic lines through careless husbandry (wandering worms explanation).
2) Starting with a silkworm line that was not carefully bred to genetic stability.
3) Imperfect genetic insertions affected by crossover.
4) Mutations
As was pointed out, mutations are rare and an unlikely cause. I doubt the 'wandering worms' explanation because silkworms mostly wander to find food and they should have ample food at all times. A mixup could only kindly be identified as 'genetic drift.' We cannot be certain, but I tend to favor explanation number 3. Others are welcome to their hypotheses as well.