why might the repair of pyrimidine dimers be faster and more efficient on the coding strand of transcriptionally active genes than on other parts of the genome?