Python Programming for Biology: Bioinformatics and Beyond

Whether to align protein or DNA

If  you’re  looking  at  the  coding  regions  of  genes  and  their  resulting  proteins,  when

approaching a particular sequence alignment problem, there is often the option of aligning

nucleotide sequences or protein sequences. In general, which kind of alignment to do will

be governed by the specific purpose of the investigation. For example, if you are studying

the  spread  of  an  influenza  virus  through  human  populations,  then  because  the  rates  of

nucleotide  change  in  the  virus  genome  are  high,  and  because  you  wish  to  study  every

genetic  change  in  fine  detail,  aligning  nucleotide  sequences  would  be  the  best  option.

Conversely,  if  you  are  interested  in  protein  structure,  maybe  by  doing  comparative

modelling,  then  using  nucleotide  sequences  would  introduce  unnecessary  error  and  you

would always use the protein sequence, albeit possibly translated from the DNA. In such

---

circumstances,  because  the  nucleotides  can  have  one  of  four  bases,  the  chances  of  a

random  match  are  quite  high  compared  to  an  amino  acid  which  is  one  of  20;  i.e.  the

chances of spurious matches in protein sequences are much smaller. Also worth noting is

the fact that some DNA changes have no effect on the protein at all, and thus are irrelevant

for many questions arising from protein sequence. Lastly, when dealing with amino acids,

we almost always calculate a score of an alignment based upon the degree to which they

match, rather than just saying how many are identical, which is often the case for DNA.

8

In general, this is possible because the chemical structure of amino acids allows you to say

how similar they are, for example, in terms of size and charge or in the ability to form a

loop.  This  can  reveal  relationships,  e.g.  conservation  caused  by  the  protein’s  structure,

which would otherwise not be visible by considering simple matches alone.