Effects of Varying Sequence Similarity on the Frequency of Repeat Deletion during Reverse Transcription of a Human Immunodeficiency Virus Type 1 Vector