Ancestral Types of Milk Protein
To examine species-specific compositional differences
in the 13 mammals, we reconstructed the ancestral proteins
of each of the milk proteins by first aligning the
orthologs of each of the 9 milk proteins using ClustalW
(Thompson et al., 1994). This was followed by a maximum
likelihood reconstruction using codeml from the
paml package (Yang, 2007). To avoid biased ancestral
sequences, we manually curated the alignments to remove
the sections of the alignments that appeared to
be a species-specific gain of a large region of protein
(an example of this would be β-CN, which has a long
insertion in rodents). Standard ancestral reconstruction
is performed in such a way that it incorporates all AA
positions, even those that are found in only one species.
To avoid this, we manually removed positions that were
only found in one species and not in the others from the
ancestral sequences.
For α-LA, we removed positions 14 and 15 (NH),
which are only found in opossum. For the same reason,
we removed position 111 (Q). We also removed positions
127 to 143 (GAPALVVPALNSETPVP) in the rat
because this region is unique to that species. For β-CN,
we removed positions 46 to 49 (LPTT) and 114 to 117
(PQQN), which are only found in platypus.