EXAMPLE ligand.dpf file.  Notes appear below.
Written by Marty Pagel
Reviewed by Karl Scheidt.

types	CNOH			# atom type names
fld	cruzdok_maps.fld	# grid data file
map	cruzdok_C.map		# C-atomic affinity map
map	cruzdok_N.map		# N-atomic affinity map
map	cruzdok_O.map		# O-atomic affinity map
map	cruzdok_H.map		# H-atomic affinity map
map	cruzdok_e.map		# electrostatics map

move	lig.pdbq		# small molecule
about	-15.475 -4.374 12.172	# small molecule center
tran0	random          	# initial coordinates/A
quat0	random     		# initial quaternion:unit-vector(qx,qy,qz);angle/deg(qw)
ndihe   11                      # numbe of dihedral angles to change during the run
dihe0   random                  #initial dihedrals/deg
tstep	0.05			# translation step/A
qstep	1.			# quaternion step/deg
dstep	1.			# torsion step/deg
trnrf	1.			# trans reduction factor/per cycle
quarf	1.			# quat reduction factor/per cycle
dihrf	1.			# tors reduction factor/per cycle

intnbp	1272653.000  1127.684  12  6 	# C-C internal energy non-bond parameters/Cn,Cm,n,m
intnbp	 610155.100   783.345  12  6 	# C-N internal energy non-bond parameters/Cn,Cm,n,m
intnbp	 588883.800   633.754  12  6 	# C-O internal energy non-bond parameters/Cn,Cm,n,m
intnbp	  88604.240   226.910  12  6 	# C-H internal energy non-bond parameters/Cn,Cm,n,m
intnbp	 266862.200   546.765  12  6 	# N-N internal energy non-bond parameters/Cn,Cm,n,m
intnbp	 249961.400   445.918  12  6 	# N-O internal energy non-bond parameters/Cn,Cm,n,m
intnbp	  39093.660   155.983  12  6 	# N-H internal energy non-bond parameters/Cn,Cm,n,m
intnbp	 230584.400   368.677  12  6 	# O-O internal energy non-bond parameters/Cn,Cm,n,m
intnbp	  38919.640   124.049  12  6 	# O-H internal energy non-bond parameters/Cn,Cm,n,m
intnbp	   1908.578    46.738  12  6 	# H-H internal energy non-bond parameters/Cn,Cm,n,m

rt0	500.		# initial RT
rtrf	0.95		# RT reduction factor/per cycle

runs	1		# number of runs 
cycles	355		# number of cycles  
accs	300000		# steps accepted
rejs	300000		# steps rejected
select	m		# minimum or last

outlev	2		# diagnostic output level

rmstol	0.5		# cluster tolerance/A

trjfrq	100000		# trajectory frequency
trjbeg	1		# start trj output at cycle
trjend	355		# end trj output at cycle
trjout	lig.trj		# trajectory file
trjsel	A		# A=acc only;E=either acc or rej

extnrg	1000.		# external grid energy

Notes about this file:
1.  Each run produces a separate docked ligand.
    You need more than 1 run if you perform culstering.
2.  Each run consists of a number of cycles (you can set
    the number of cycles by setting "cycles  ###" above).  
3.  At the start of each cycle, the coordinates of the 
    ligand are randomly changed.  If the resulting structure
    is LOWER in energy, this step is always accepted.  If the 
    resulting structure is HIGHER in energy, this step 
    may be accepted based upon the Boltzman probability
    distribution (see manual).  In other words, when the
    temperature is high, many steps that are higher in energy 
    than the previous step are accepted; when the temperature 
    is low, few steps that are higher in energy than the previous 
    step are accepted.  When the total number of accepted 
    or rejected steps meets the 'accs' or 'rejs' numbers you
    type in above, then that cycle is finished.
4.  When a new cycle begins, the minimum or last structure from
    the previous cycle is used (depending if 'select' is 'm' 
    or 'l').  The temperature for the cycle is set to:
    (previous cycle temp)(rtrf) 
5.  Structures that have and RMSD < rmstol when they are 
    superimposed will be 'binned' together (e.g., the program
    will 'cluster' these structures and tell you that they 
    are part fo the same family of structures.)
6.  A trajectory frame is saved every trjfrq steps.  Make sure
    that the 'trjout' file exists and is empty before starting 
    autodock.
7.  Don't forget to include ndihe and dihe0, or your totatable bonds
    won't rotate.
8.  While our results are preliminary, it appears that the calculation
    rejects all steps once the temperature is < 2 or 3K.  Therefore, 
    the number of cycles should be set to:
    cycles = ln(2/rt0) / ln(rtrf)
    It also appears that rt0 = 500 is OK.
    Finally, we think the rate of cooling should be as slow as 
    reasonalbly possible.  We set rtrf=0.99 and accs=rejs=80000.
    A typical calculation with these parameters and run=1 is 
    about 6 hours on a Indy 180 MHz R5000SC.