In viruses, bacteria etc., the genomes are small and consists of very less number of nucleotide.
But higher the animal in evolutionary rank, genomic complexity is higher.
So changing of some genes by cut paste method (genetic engineering) can cause change in surface protein structure of virus or bacteria.
But in human, generally a single Gene do not cause physical expression of any character. Many genes of same chromosome or even other chromosome control physical expression of one character.
So change of any character by Gene substitution is not easy in human.
In plants we know that resistance to a number of fungal, bacterial and viral pathogens is conferred by single genes. The classic is Verticillium wilt in tomatoes. This particular gene is inherited by classic Mendelian principals and so breeding for resistant tomatoes is a simple matter. You simply discard the plants that do not carry the appropriate gene combinations. In humans, a classic case is malaria resistance by the sickle cell trait in Africans. This is inherited by a classic Mendelian gene; 1/2 of the offspring of two heterozygotes are resistant to malaria (a good thing) but 1/4 are susceptible to the disease and the remaining 1/4 have the sickle cell trait and typically die of this particular anemia. In an evolutionary sense this is reasonable but for families to have a 1/4 chance that their child will die a painful death is unacceptable - we simply cannot "discard" or restrict the breeding of human beings with incorrect genetic combinations. Hypothetically, new technology that could transfer a gene for disease resistance to a rhinovirus or even a disabled herpes virus used to infect humans and integrate the resistant gene into the patients' genome is possible. However, we have not identified single genes that are responsible for disease resistance in humans. Typically disease resistance in humans is a multi-gene trait that is inherited quantitatively and not in a Mendelian manner. A good idea but we are still quite a long way from implementation.