ConnectedCloud

Overview

ConnectedCloud grows a cloud of objects that are "connected" through single-linkage clustering. Two objects are "connected" if they are within the user-defined cutoff, radius. It starts with a selection, and grows until no more objects are withing radius Angstroms of the cloud.

Let's say we want to find the cloud of docking guesses (blue lines) that are connected to the lowest energy docking pose (red sticks). I call,

connectedCloud lig1, lig*, radius=1, doPrint=True

and get

• 

  Find the cloud growing from the red sticks. Consider all blue lines for inclusion.

• 

  Those in the cloud are shown in red. Notice the small blue cloud above the red was not selected. At no point were those atoms close enough to the currently growing cloud, to be considered.

Syntax

connectedCloud origSel, subSel, [radius=1, [doPrint=True]]

• origSel -- the original selection to grow from
• subSel -- only consider these objects when growing
• radius -- defines "connectedness". Anything <= radius Angstroms from the growing cloud and in subSel is considered
• doPrint -- print some misc. info

Examples

# Find all objects in the set "lig*" (all ligands from a docking run, for example) that are
# in the connected cloud of "connectedness" set to 1 Ang.  Print the number of atoms in the
# selection and the surface area of the selection (if it were an object).
connectedCloud lig427, lig*, radius=1, doPrint=True

# select everything within 0.85 of lig2467 and don't print anything
connectedCloud lig2467, *, radius=0.85, doPrint=False

The Code

# -*- coding: utf-8 -*-
def connectedCloud( origSel, subSel, radius=1, doPrint=True ):
	"""
	connectedCloud -- find all atoms, by object in subSel connected to
		origSel by radius Ang.  In other words, start from origSel
		and find all objects <=radius Ang. from it & add that to the
		result.  Once nothing more is added, you have your cloud of
		points.

	origSel,
		the selection around which we start expanding.

	subSel,
		the subset of objects to consider

	radius,
		the maximum radius in which to consider; anything >radius
		Ang. from any objet in the current selection is ignored (unless
		it is added later by some other path).

	doPrint
		print some values before returning the selection name

	Author:
		Jason Vertrees, 2009.
	
	"""
	cName = "__tempCloud__"
	cmd.select(cName, origSel)

	# we keep track of cloud size based on the number of atoms in the selection; oldCount
	# is the previous iteration and curCount is after we add everything for this iteration.
	oldCount=0
	curCount=cmd.count_atoms( cName )

	# flag to stop at 1000 iterations. if something goes wrong we don't want to hijack the computer.
	flag=0

	# while something was added in this iteration
	while oldCount != curCount:
		flag += 1
		# grow current cloud
		cmd.select( cName, "bo. " + subSel + " within 1 of " + cName )
		# update atom counts in the cloud
		oldCount=curCount
		curCount=cmd.count_atoms(cName)
		
		if ( flag > 1000 ):
			print "Warning: I gave up.  I grew the cloud 1000 times and it kept growing."
			print "Warning: If this is valid (weird), then increases the 'flag' in the source"
			print "Warning: code to something >1000."
			break

	# create the object for getting its surface area
	cmd.create("connectedCloud", cName)

	# print cloud info
	if doPrint:
		print "Number of atoms in cloud:", str(cmd.count_atoms(cName))
		print "Surface area of cloud:", str(cmd.get_area("connectedCloud"))

	# delete our work 
	cmd.delete("connectedCloud")
	# rename and return
	cmd.set_name(cName, "connectedCloud")
	return "connectedCloud"

cmd.extend("connectedCloud", connectedCloud)