Figure 1

Nature of HERV-KX sequences. (A) Similarity of 642 HML-2 cDNA sequences and their best matching HML-2 loci in the human genome. The majority of cDNAs displayed between zero and a few differences to the best match, and were thus assignable to specific HML-2 loci. A minority of cDNAs displayed a greater number of dissimilarities to the best match and were thus not assignable with confidence to specific HML-2 loci. HERV-KX sequences were defined as displaying 18 or more nucleotide differences to the best matching HML-2 locus. (B) Sequence divergence of HERV-KX sequences in comparison to selected HML-2 reference sequences from the human genome, depicted as a neighbour joining-tree of the absolute number of nucleotide differences between sequences. For the sake of clarity, phylogenetically more distant HML-2 reference sequences were not included in the tree, as they were less similar to HERV-KX sequences than the reference sequences included in the tree. Proviral reference sequences are given as "xx_xxx" (see text). Positions with gaps were excluded in pairwise sequence comparisons.