NAME

mkd3d - generate descriptor for closed circular 3DNA model

SYNOPSIS

mkd3d [ options ] -n N_bp -c conf -h helf -d desf

mkd3d [ options ] -n N_bp -c conf -s seqf -d desf

mkd3d [ options ] -n N_bp -c conf -H vhel -d desf

DESCRIPTION

Generate a text descriptor file for the 3DNA reduced representation model of DNA named desf containing N_bp base pairs, using as input the constants file conf and the helical file helf (in the first form of the command) or the sequence file seqf (in the second form of the command) or the helical file vhel in the old AUGUR format (in the third form of the command). The files helf or vhel or seqf must contain parameters for at least N_bp base pairs. The use of the old AUGUR format helical file is discouraged and therefore the last form of the command should not be used.

The generated force field takes its equilibrium dimensions from the helical files or when a sequence file is specified instead these dimensions are predicted from the base sequence using the Bolshoy rules. The force constants are taken from the constants file.

The options are:

-open: Generate a linear (open) model (the two ends of the DNA are not linked) instead of the preset behavior of closing into a circular model (the two ends of the DNA are linked to form a closed circle)

-x C: Ignore the first C parameters in the helical file. The i-th base pair of the model will receive the helical parameters of the C+i-th base pair of the helical file.

-z Z: Generate a double well potential for the twist of the bases named in Z. The double well potential has minimum energies at ±t_o degrees where t_o is the equilibrium twist from the iton record of the constants file. Z is a continuous string (no spaces) containing numbers that are interpreted as follows: "a,bb,c-dd,e(f)ggg" where a, bb, c, dd, e, f, ggg are positive whole numbers, the string expands to mean the following base pairs: a, bb, c, c+1, c+2 to dd, e, e+f, e+2f to ggg.

-aroll A: generate anisotropic roll bending. The preset action is to set A to one in which case the roll angle is represented by a simple angle function with a force constant k_roll and equilibrium roll angle q_o. When -aroll is specified, the bending in the roll direction is represented by a piecewise power series: for roll angle 0 < q <= q_o the potential function is k_roll[q - q_o]²/A and for roll angle q_o < q <= 180 degrees the potential function is Ak_roll[q - q_o]². This is an asymmetrical function with the minimum at q_o.

In the 3DNA model, each base pair of the DNA is represented by three particles. The first is located in the major groove and is named "FRON", the second is located on one of the back bones and is called "LEFT" and the last particles is located in the center of the base pair and is called "CENT". Each group of three atoms may be assigned to one of four groups named "A", "C", "G" or "T" and different physical properties may be assigned according to the group. (These names appear in the name record of the generated descriptor.)

A text descriptor is generated rather than the binary version. This is to allow modification of the descriptor. One would almost certainly want to add notes to the descriptor (see the note record of the descriptor file format). Once a final descriptor is obtained, convert it to binary. See des.

EXAMPLES

mkd3d -n 360 -c 3dna.cnt -h circ360.hel -d circ360.dx

Create a text descriptor named circ360.dx for a 360 base pair 3DNA model using the force constants from 3dna.cnt and the equilibrium dimensions from the helical file circ360.hel.

mkd3d -s c1500.seq -d c1500.dx -n 1500 -open -c 3dna.cnt

Create a text descriptor named c1500.dx for a 1500 base pair 3DNA model that is open (linear), using the force constants from 3dna.cnt and the equilibrium dimensions predicted from the base sequence in c1500.seq using the Bolshoy rules.

FILES

The constants file (-c) has the following format:

The file starts with the header line

CON1

1000

Nclass

The file is identified by the keyword CON1 as a constants file and the following number is the version number which must be 1000 and the last number N_class is the maximum number of classes which must be a number larger than the number of bonds, angles, improper torsions and so on that will follow in the file. These records are identified by keywords and can appear in any order. Comments can be placed between the character # and the end of line and may appear between records but not within the records. The records must conform to the 3DNA model so they have a rather strict layout order.

The bond record is laid out as follows:

bond	Nbond
	kbij	boij
	...	...

Following the keyword bond is the number of bond records that will follow N_bond. Each bond record consists of the force constant k_bij [erg/Ų] and the equilibrium bond length b_oij[Å]. In practice, N_bond is always 12 and the first bond record must be for the CENT_i-1-CENT_i bond (the subscript denoting the base pair number), the second for CENT_i-FRON_i and the third for CENT_i-LEFT_i. The first three records are for base type A, the next groups of three bonds are for C, G and T in that order, each group of three are for the three bonds listed here. Typically the equilibrium bond lengths are 3.4, 2.5 and 5Å respectively.

The angle record is laid out as follows:

angl	Nangl
	kqijk	qoijk
	...	...

Following the keyword angl is the number of angles that will follow N_angl. Each angle record consists of the force constant k_qijk[erg/rad²] and the equilibrium angle q_oijk[deg]. In practice, N_angl is always 20 and the first five angles are for base type A the next groups of five are for C, G and T in that order. In each group of five angles, the first angle is for CENT_i-1-CENT_i-FRON_i (the "roll" angle) the second for CENT_i-1-CENT_i-LEFT_i (the "tilt" angle) the third for FRON_i-CENT_i- LEFT_i, the fourth for LEFT_i-CENT_i-CENT_i+1 and the fifth for FRON_i-CENT_i-CENT_i+1. The equilibrium angle for the first two angles are ignored and are instead taken from the helical files or predicted from the Bolshoy rules. The last three angles are usually assigned equilibrium values of 90 degrees to keep the base plane regular.

The improper torsion record is laid out as follows:

itor	Nitor
	ktijkl	toijkl
	...	...

Following the keyword itor is the number of improper torsion records N_itor. Each improper torsion record consists of the force constant k_tijkl[erg/rad²] and the equilibrium torsion angle t_oijkl[deg]. In practice, N_itor is always 4 and the torsion angle is defined by the atoms LEFT_i-1-CENT_i-1-CENT_i-LEFT_i (the "twist" angle) and the four torsions are defined for base types A, C, G and T in that order. The equilibrium torsion angle is ignored and is instead taken from the helical file or generated from the base sequence using the Bolshoy rules.

The torsion power series record is laid out as follows:

iton	Niton
	Eijkl	toijkl
	...	...

Following the keyword iton is the number of torsion power series terms N_iton. The records consist of an energy barrier E_ijkl [erg] and the equilibrium angle t_oijkl. This record is used only when the -z option is active. In practice N_iton is always 4 and the four records are for base types A, C, G and T in that order. A power series in torsion will be used to represent the following function: E_ijkl[t_ijkl²-t_oijkl²]²/t_oijkl⁴. which is a symmetrical function with zeros at torsion angles ±t_oijkl and a barrier of E_ijkl at zero torsion.

The non-bond record is laid out as follows:

nbn+	Nnb	n	Di-j	Dj-j
	knij	doij
	...	...

Following the keyword nbn+ is the number of non-bond records N_nb, the power n and the exclusion parameters D_i-j and D_j-j. In the 3DNA model, each base pair, represented by three particles, are connected through CENT to form a long chain. Only CENT has any bulk and advantage is taken of the fact that DNA is locally stiff. This means that the non-bond interactions need to be evaluated for only a small number of atoms. In addition to ignoring two thirds of the atoms (the FRON and LEFT atoms), for each CENT atom, non-bond interactions are considered only for other CENT atoms that are separated by at least D_i-j bonds and then only every D_j-j -th CENT atoms thereafter. Each non-bond record consists of the force constant k_nij[erg/Ų] and the contact exclusion distance d_oij[Å]. In practice, k_nij is always 4 and the non-bond records are to be given for A, C, G and T in that order.

The mass record is laid out as follows:

mass	Nmass
	Mi
	...

Following the keyword mass is the number of mass records N_mass. The masses M_i are in units of amu. In practice, N_mass is always 12 and four groups of masses are to be given the groups being for the base types A, C, G and T in that order. Each group of masses are for FRON, LEFT and CENT in that order.

The helical file format (-h) has the following layout:

The first line of the file must contain the following five words, they can be separated by any number of spaces but they must be in exact order.

Base

Rise[ang]

Twist[deg]

Roll[deg]

Tilt[deg]

The helical parameters are then assumed to follow and are read as follows:

Bi	Rise	Twist	Roll	Tilt
...	...	...	...	...

where Bi is a string that starts with a single character denoting the base symbol followed without an intervening space by the base pair number, for example: G1501 for the base symbol for guanine and a base pair number 1501. Only the following base symbols are recognized: A, C, G and T. This is followed by four numbers that are the rise in Å and the twist, roll and tilt angles in degrees.

The sequence file format (-s) has the following layout.

The sequence file may contain any number of text lines which will be ignored until the occurrence of two contiguous periods.

text

..

Once the two periods appear together, the comment is assumed to have ended and the rest of the file will contain the base sequence.

Columns 1 to 10 are ignored. (This may contain a sequence number to keep track of the lines.) The base sequence starts on column 11 and the base symbols are separated by any number of white spaces.

1234567890	A CCGT TT AAAA GCGGG...
	...

The only base symbols that are allowed to appear are A, C, G and T. The base sequence is then used to generate the helical parameters using the Bolshoy, McNamara and Trifonov rules. The helical parameters are then used as equilibrium dimensions in the generated model.

The old AUGUR helical file (-H) is laid out as follows:

The file starts with the header line

HEL1

1000

The file identifier is the string HEL1 which must start on column one of line one. This is followed by the file version number which must be 1000.

This is followed by any number of lines until one that has the format:

Helix Parameters (degrees), bp:bp[A]=

R

This line must start on column two (with a space on column one) and be exactly as shown above; the only variable is R the rise in Å.

The appearance of this line signals the start of the data lines which will be read as:

Bi	Tw	Ro	Ti	Bi	Tw	Ro	Ti
...	...	...	...	...	...	...	...

where Bi is a string starting with a single letter denoting the base symbol B (e.g., A,C,G,T) and followed by the number i, for the i-th base pair, e.g. A301 for adenine base pair 301. Tw, Ro and Ti are the twist, roll and tilt angles respectively all expressed in degrees. A pair of base symbols and helical parameters are expected on each line. The first Bi starts on column 5 and the second starts on column 45. The angles can be placed freely within these limits separated by spaces. The helical parameters need not appear in proper sequence. The parameters are assigned to the i-th base pair as read from this file. If the same base pair appears more than once, the parameters last read will be in effect. If parameters are not defined for a base pair, a descriptor is still produced but the file will not be of any use.

SEE ALSO

mkd3c, xmd3, des, the descriptor file format.

DIAGNOSTICS

BUGS

NOTES

mkd3d is one of three programs specifically designed for the 3DNA model. In this model, the base pairs of a DNA double helix is represented by only three masses. This reduces the size of the problem by one to two orders of magnitude allowing large pieces of DNA to be studied.