Source of the materials: Biopython cookbook (adapted) Status: Draft
If you haven’t spent a lot of time programming in Python, many questions and problems that come up in using Biopython are often related to Python itself. This section tries to present some ideas and code that come up often (at least for us!) while using the Biopython libraries. If you have any suggestions for useful pointers that could go here, please contribute!
Handles are mentioned quite frequently throughout this documentation, and are also fairly confusing (at least to me!). Basically, you can think of a handle as being a “wrapper” around text information.
Handles provide (at least) two benefits over plain text information:
They provide a standard way to deal with information stored in different ways. The text information can be in a file, or in a string stored in memory, or the output from a command line program, or at some remote website, but the handle provides a common way of dealing with information in all of these formats.
They allow text information to be read incrementally, instead of all at once. This is really important when you are dealing with huge text files which would use up all of your memory if you had to load them all.
Handles can deal with text information that is being read (e. g. reading
from a file) or written (e. g. writing information to a file). In the
case of a “read” handle, commonly used functions are read()
, which
reads the entire text information from the handle, and readline()
,
which reads information one line at a time. For “write” handles, the
function write()
is regularly used.
The most common usage for handles is reading information from a file,
which is done using the built-in Python function open
. Here, we open a
handle to the file m_cold.fasta (also
available online
here):
In [5]:
handle = open("data/m_cold.fasta", "r")
handle.readline()
Out[5]:
Handles are regularly used in Biopython for passing information to
parsers. For example, since Biopython 1.54 the main functions in
Bio.SeqIO
and Bio.AlignIO
have allowed you to use a filename instead
of a handle:
In [7]:
from Bio import SeqIO
for record in SeqIO.parse("data/m_cold.fasta", "fasta"):
print(record.id, len(record))
On older versions of Biopython you had to use a handle, e.g.
In [9]:
from Bio import SeqIO
handle = open("data/m_cold.fasta", "r")
for record in SeqIO.parse(handle, "fasta"):
print(record.id, len(record))
handle.close()
This pattern is still useful - for example suppose you have a gzip compressed FASTA file you want to parse:
import gzip
from Bio import SeqIO
handle = gzip.open("m_cold.fasta.gz")
for record in SeqIO.parse(handle, "fasta"):
print(record.id, len(record))
handle.close()
See Section [sec:SeqIO_compressed] for more examples like this, including reading bzip2 compressed files.
One useful thing is to be able to turn information contained in a string
into a handle. The following example shows how to do this using
cStringIO
from the Python standard library:
In [1]:
my_info = 'A string\n with multiple lines.'
print(my_info)
In [4]:
from io import StringIO
my_info_handle = StringIO(my_info)
first_line = my_info_handle.readline()
print(first_line)
In [10]:
second_line = my_info_handle.readline()
print(second_line)
In [ ]: