Python Programming for Biology: Bioinformatics and Beyond

Quick database searches

As  a  final  thought  to  this  chapter  we  acknowledge  another  kind  of  alignment  that  is

routinely performed. This is to have a query sequence and compare it with a large database

of other sequences to find alignment matches from a potentially vast number of targets.

What  we  have  been  discussing  so  far  are  rigorous  alignments  of  limited  numbers  of

sequence  pairs.  The  dynamic  programming  method  and  its  various  derivatives  will  give

you the best alignments, but it has the disadvantage that it is prohibitively slow if you are

aligning a query sequence with a database containing thousands (e.g. the human genome)

or even millions of sequences (e.g. the whole EMBL or NCBI database). When searching

such databases, the approach is to use query methods that can take clever short-cuts, i.e. it

uses rules that are generally true in order to quickly remove a large number of sequences

from consideration, without having to do full sequence alignments with every member of

the database. The most famous of these methods is the BLAST routine.

17

Although such

heuristic methods have opened up a vast array of convenient sequence information, strictly

speaking they can miss things that a more rigorous approach would not.

Below we give an example of how we can run BLAST searches on sequence databases

using Python. The objective here is not to rewrite the BLAST program itself, but rather to

wrap  the  external  program  in  code  and  use  it  as  if  it  were  a  Python  module  which  can

interface with all of our other code. This naturally has the advantage of using an efficient,

tried and tested program but still being able to work with Python. Thus what we will do is

create a function that makes the required input, runs the external executable (i.e. the actual

BLAST program, which must already be installed on the system) and interprets the output,

collecting the data back into the Python data structures.