Wed, Oct 23, 2013 - More Python; Annotation; Scripting
######################################################

Python homework
===============

HW3 IPython Notebook solutions set: http://nbviewer.ipython.org/7120119

Prokka and annotation evaluation
================================

What does Prokka do?  Basically annotation: finds non-coding and
protein-coding genes; provides provisional names.

Re HW3, how did we use Prokka here? Was there a clear "winner" in terms of k value? What further
questions might we ask?

BLAST output and parsing with Python
------------------------------------

Working with BLAST output: basic export from BLAST to CSV,

   https://github.com/ngs-docs/ngs-scripts/blob/master/blast/blast-to-csv.py

and also:

   https://github.com/ngs-docs/ngs-scripts/blob/master/blast/blast-to-ortho-csv.py

Automating stuff; scripting
===========================

One (a major) reason why the command line has been so successful is that
it's pretty easy to write software that can mixed, matched, and placed in
automated pipelines at the command line.

See::

      for i in {19..51..2}
      do
         /mnt/prokka-1.7/bin/prokka ecoli-k${i}.fa --outdir k${i} --prefix k${i}
	 # other commands involving ${i}
      done

Note, this is a *completely* separate language from Python! This is shell
programming, not Python programming.  It is possible to mix and match them
but as a rule I try to write *complicated scripts* in Python and reserve
shell programming for when I'm using the command line interactively.  This
works for me but Your Mileage May Vary (YMMV).

Shell scripts
-------------

"Scripts" can be written in many different languages: shell and Python are
the two you have experience with, but R also.

The idea is that you just put the commands you want to run into a file
and then run that file with the right command: 'bash' for shell scripts,
'python' for 

To try it, do::

   echo "echo hello, world" > hello.sh
   bash hello.sh

and::

   echo "print 'hello, world'" > hello.py
   python hello.py

----

Creating and editing these files is tricky, however.  Need to use a text
editor: can use 'pico' locally
(or other, more complicated editors); or TextEdit/TextWrangler remotely,
then copy via Dropbox.

----

GitHub
======

Gists: place to store random "stuff".

Github more generally: version controlled.

Go to https://github.com/ngs-docs/ngs-scripts/; make sure you're logged in;
click "fork".  Click edit.  Make some changes.  Save, with commit.