Shortcut: SKETCHMAP

Module	dimred
Description	Usage
Construct a sketch map projection of the input data
output value	type
the sketch-map projection of the input points	matrix

Details and examples

Construct a sketch map projection of the input data

This shortcut allows you to construct a sketch-map projection from a trajectory. The sketch-map algorithm is introduced and examples of how it is used are given in the papers cited below.

The following input illustrates how to run a sketch-map calculation with PLUMED:

Click on the labels of the actions for more information on what each action computes

d1The DISTANCE action with label d1 calculates the following quantities:
 Quantity   
 Type   
 Description  
d1
scalar
the DISTANCE between this pair of atoms: DISTANCECalculate the distance/s between pairs of atoms. More details ATOMSthe pair of atom that we are calculating the distance between=1,2
d2The DISTANCE action with label d2 calculates the following quantities:
 Quantity   
 Type   
 Description  
d2
scalar
the DISTANCE between this pair of atoms: DISTANCECalculate the distance/s between pairs of atoms. More details ATOMSthe pair of atom that we are calculating the distance between=3,4
d3The DISTANCE action with label d3 calculates the following quantities:
 Quantity   
 Type   
 Description  
d3
scalar
the DISTANCE between this pair of atoms: DISTANCECalculate the distance/s between pairs of atoms. More details ATOMSthe pair of atom that we are calculating the distance between=5,6

ffThe COLLECT_FRAMES action with label ff calculates the following quantities:
 Quantity   
 Type   
 Description  
ff_data
matrix
the data that is being collected by this action
ff_logweights
vector
the logarithms of the weights of the data points: COLLECT_FRAMESCollect atomic positions or argument values from the trajectory for later analysis This action is a shortcut and it has hidden defaults. More details STRIDE the frequency with which data should be stored for analysis=1 ARGthe labels of the values whose time series you would like to collect for later analysis=d1,d2,d3
The COLLECT_FRAMES action with label ff calculates the following quantities:
 Quantity   
 Description  
ff.data
the data that is being collected by this action
ff.logweights
the logarithms of the weights of the data pointsff: COLLECT_FRAMESCollect atomic positions or argument values from the trajectory for later analysis This action is a shortcut and uses the defaults shown here. More details STRIDE the frequency with which data should be stored for analysis=1 ARGthe labels of the values whose time series you would like to collect for later analysis=d1,d2,d3  CLEAR the frequency with which data should all be deleted and restarted=0
# PLUMED interprets the command:
# ff: COLLECT_FRAMES STRIDE=1 ARG=d1,d2,d3
# as follows (Click the red comment above to revert to the short version of the input):
ff_d1The COLLECT action with label ff_d1 calculates the following quantities:
 Quantity   
 Type   
 Description  
ff_d1
vector
the time series for the input quantity: COLLECTCollect data from the trajectory for later analysis More details ARGthe label of the value whose time series is being stored for later analysis=d1 STRIDE the frequency with which the data should be collected and added to the quantity being averaged=1 CLEAR the frequency with which to clear all the accumulated data=0
ff_d2The COLLECT action with label ff_d2 calculates the following quantities:
 Quantity   
 Type   
 Description  
ff_d2
vector
the time series for the input quantity: COLLECTCollect data from the trajectory for later analysis More details ARGthe label of the value whose time series is being stored for later analysis=d2 STRIDE the frequency with which the data should be collected and added to the quantity being averaged=1 CLEAR the frequency with which to clear all the accumulated data=0
ff_d3The COLLECT action with label ff_d3 calculates the following quantities:
 Quantity   
 Type   
 Description  
ff_d3
vector
the time series for the input quantity: COLLECTCollect data from the trajectory for later analysis More details ARGthe label of the value whose time series is being stored for later analysis=d3 STRIDE the frequency with which the data should be collected and added to the quantity being averaged=1 CLEAR the frequency with which to clear all the accumulated data=0
ff_dataThe VSTACK action with label ff_data calculates the following quantities:
 Quantity   
 Type   
 Description  
ff_data
matrix
a matrix that contains the input vectors in its columns: VSTACKCreate a matrix by stacking vectors together More details ARGthe values that you would like to stack together to construct the output matrix=ff_d1,ff_d2,ff_d3
ff_cweightThe CONSTANT action with label ff_cweight calculates the following quantities:
 Quantity   
 Type   
 Description  
ff_cweight
scalar
the constant value that was read from the plumed input: CONSTANTCreate a constant value that can be passed to actions More details VALUEthe single number that you would like to store=0
ff_logweightsThe COLLECT action with label ff_logweights calculates the following quantities:
 Quantity   
 Type   
 Description  
ff_logweights
vector
the time series for the input quantity: COLLECTCollect data from the trajectory for later analysis More details ARGthe label of the value whose time series is being stored for later analysis=ff_cweight STRIDE the frequency with which the data should be collected and added to the quantity being averaged=1 CLEAR the frequency with which to clear all the accumulated data=0
ff_oneThe CONSTANT action with label ff_one calculates the following quantities:
 Quantity   
 Type   
 Description  
ff_one
scalar
the constant value that was read from the plumed input: CONSTANTCreate a constant value that can be passed to actions More details VALUEthe single number that you would like to store=1
ff_onesThe COLLECT action with label ff_ones calculates the following quantities:
 Quantity   
 Type   
 Description  
ff_ones
vector
the time series for the input quantity: COLLECTCollect data from the trajectory for later analysis More details ARGthe label of the value whose time series is being stored for later analysis=ff_one STRIDE the frequency with which the data should be collected and added to the quantity being averaged=1 CLEAR the frequency with which to clear all the accumulated data=0
# --- End of included input --- lweThe CUSTOM action with label lwe calculates the following quantities:
 Quantity   
 Type   
 Description  
lwe
vector
the vector obtained by doing an element-wise application of an arbitrary function to the input vectors: CUSTOMCalculate a combination of variables using a custom expression. More details ARGthe values input to this function=ff_logweights FUNCthe function you wish to evaluate=exp(x) PERIODICif the output of your function is periodic then you should specify the periodicity of the function=NO
smapThe SKETCHMAP action with label smap calculates the following quantities:
 Quantity   
 Type   
 Description  
smap
matrix
the sketch-map projection of the input points: SKETCHMAPConstruct a sketch map projection of the input data This action is a shortcut and it has hidden defaults. More details ...
   ARGthe matrix of high dimensional coordinates that you want to project in the low dimensional space=ff NLOW_DIMnumber of low-dimensional coordinates required=2
   HIGH_DIM_FUNCTIONthe parameters of the switching function in the high dimensional space={SMAP R_0=2 A=3 B=9}
   LOW_DIM_FUNCTIONthe parameters of the switching function in the low dimensional space={SMAP R_0=2 A=2 B=2}
   WEIGHTSa vector containing the weights of the points=lwe NCYCLES The number of cycles of pointwise global optimisation that are required=3 FGRID_SIZE interpolate the grid onto this number of points -- only works in 2D=50,50
...
smap: SKETCHMAPConstruct a sketch map projection of the input data This action is a shortcut and uses the defaults shown here. More details ...
   ARGthe matrix of high dimensional coordinates that you want to project in the low dimensional space=ff NLOW_DIMnumber of low-dimensional coordinates required=2
   HIGH_DIM_FUNCTIONthe parameters of the switching function in the high dimensional space={SMAP R_0=2 A=3 B=9}
   LOW_DIM_FUNCTIONthe parameters of the switching function in the low dimensional space={SMAP R_0=2 A=2 B=2}
   WEIGHTSa vector containing the weights of the points=lwe NCYCLES The number of cycles of pointwise global optimisation that are required=3 FGRID_SIZE interpolate the grid onto this number of points -- only works in 2D=50,50
 CGTOL The tolerance for the conjugate gradient minimization that finds the projection of the landmarks=1E-6 MAXITER maximum number of optimization cycles for optimisation algorithms=1000 BUFFER grid extent for search is (max projection - minimum projection) multiplied by this value=1.1 CGRID_SIZE number of points to use in each grid direction=10,10
...
# PLUMED interprets the command:
# smap: SKETCHMAP ...
#    ARG=ff NLOW_DIM=2
#    HIGH_DIM_FUNCTION={SMAP R_0=2 A=3 B=9}
#    LOW_DIM_FUNCTION={SMAP R_0=2 A=2 B=2}
#    WEIGHTS=lwe NCYCLES=3 FGRID_SIZE=50,50
# ...
# as follows (Click the red comment above to revert to the short version of the input):
smap_mdsThe VSTACK action with label smap_mds calculates the following quantities:
 Quantity   
 Type   
 Description  
smap_mds
matrix
a matrix that contains the input vectors in its columns: CLASSICAL_MDSCreate a low-dimensional projection of a trajectory using the classical multidimensional More details ARGthe arguments that you would like to make the histogram for=ff NLOW_DIMnumber of low-dimensional coordinates required=2
smap_hdmatThe MORE_THAN action with label smap_hdmat calculates the following quantities:
 Quantity   
 Type   
 Description  
smap_hdmat
matrix
the matrix obtained by doing an element-wise application of a function that is one if the if the input is more than a threshold to the input matrix: MORE_THANUse a switching function to determine how many of the input variables are more than a certain cutoff. More details ARGthe values input to this function=smap_mds_mat SQUARED is the input quantity the square of the value that you would like to apply the switching function to SWITCHThis keyword is used if you want to employ an alternative to the continuous swiching function defined above={SMAP R_0=2 A=3 B=9}
lwe_sumThe SUM action with label lwe_sum calculates the following quantities:
 Quantity   
 Type   
 Description  
lwe_sum
scalar
the SUM of the elements in the input value: SUMCalculate the sum of the arguments More details ARGthe vector/matrix/grid whose elements shuld be added together=lwe PERIODICif the output of your function is periodic then you should specify the periodicity of the function=NO
lwe_normedThe CUSTOM action with label lwe_normed calculates the following quantities:
 Quantity   
 Type   
 Description  
lwe_normed
vector
the vector obtained by doing an element-wise application of an arbitrary function to the input vectors: CUSTOMCalculate a combination of variables using a custom expression. More details ARGthe values input to this function=lwe,lwe_sum FUNCthe function you wish to evaluate=x/y PERIODICif the output of your function is periodic then you should specify the periodicity of the function=NO
lwe_matThe OUTER_PRODUCT action with label lwe_mat calculates the following quantities:
 Quantity   
 Type   
 Description  
lwe_mat
matrix
a matrix containing the outer product of the two input vectors that was obtained using the function that was input: OUTER_PRODUCTCalculate the outer product matrix of two vectors More details ARGthe labels of the two vectors from which the outer product is being computed=lwe_normed,lwe_normed
smap_cgThe ARRANGE_POINTS action with label smap_cg calculates the following quantities:
 Quantity   
 Type   
 Description  
smap_cg.coord-1
vector
the coordinates of the points in the low dimensional space  This is the 1th of these quantities
smap_cg.coord-2
vector
the coordinates of the points in the low dimensional space  This is the 2th of these quantities: ARRANGE_POINTSArrange points in a low dimensional space so that the (transformed) distances between points in the low dimensional space match the dissimilarities provided in an input matrix. More details ARGthe initial positions for the projections=smap_mds-1,smap_mds-2 MINTYPE the method to use for the minimisation=conjgrad TARGET1the matrix of target quantities that you would like to match=smap_hdmat FUNC1a function that is applied on the distances between the points in the low dimensional space={SMAP R_0=2 A=2 B=2} WEIGHTS1the matrix with the weights of the target quantities=lwe_mat CGTOL the tolerance for the conjugate gradient minimization=1E-6
smap_apThe ARRANGE_POINTS action with label smap_ap calculates the following quantities:
 Quantity   
 Type   
 Description  
smap_ap.coord-1
vector
the coordinates of the points in the low dimensional space  This is the 1th of these quantities
smap_ap.coord-2
vector
the coordinates of the points in the low dimensional space  This is the 2th of these quantities: ARRANGE_POINTSArrange points in a low dimensional space so that the (transformed) distances between points in the low dimensional space match the dissimilarities provided in an input matrix. More details ARGthe initial positions for the projections=smap_cg.coord-1,smap_cg.coord-2 NCYCLES the number of cycles of global optimization to attempt=3 BUFFER grid extent for search is (max projection - minimum projection) multiplied by this value=1.1 FGRID_SIZE interpolate the grid onto this number of points -- only works in 2D=50,50 CGRID_SIZE number of points to use in each grid direction=10,10 MINTYPE the method to use for the minimisation=pointwise TARGET1the matrix of target quantities that you would like to match=smap_hdmat FUNC1a function that is applied on the distances between the points in the low dimensional space={SMAP R_0=2 A=2 B=2} WEIGHTS1the matrix with the weights of the target quantities=lwe_mat CGTOL the tolerance for the conjugate gradient minimization=1E-6
smapThe VSTACK action with label smap calculates the following quantities:
 Quantity   
 Type   
 Description  
smap
matrix
a matrix that contains the input vectors in its columns: VSTACKCreate a matrix by stacking vectors together More details ARGthe values that you would like to stack together to construct the output matrix=smap_ap.coord-1,smap_ap.coord-2
# --- End of included input --- 
DUMPVECTORPrint a vector to a file More details ARGthe labels of vectors/matrices that should be output in the file=smap.*,lwe FILE the file on which to write the vetors=smap

The sketch-map projection is constructed from all the data in the input trajectory here and is output at the end of the simulation. Dissimilarities between the trajectory frames are calculated based on how much the three distances that are used in the input to the COLLECT_FRAMES object here have changed. However, if you want to use RMSD distances to compute these dissimilarities instead you can use an input like the one shown below:

Click on the labels of the actions for more information on what each action computes

ffThe COLLECT_FRAMES action with label ff calculates the following quantities:
 Quantity   
 Type   
 Description  
ff_data
matrix
the data that is being collected by this action
ff_logweights
vector
the logarithms of the weights of the data points: COLLECT_FRAMESCollect atomic positions or argument values from the trajectory for later analysis This action is a shortcut and it has hidden defaults. More details STRIDE the frequency with which data should be stored for analysis=1 ATOMSlist of atomic positions that you would like to collect and store for later analysis=1-256
The COLLECT_FRAMES action with label ff calculates the following quantities:
 Quantity   
 Description  
ff.data
the data that is being collected by this action
ff.logweights
the logarithms of the weights of the data pointsff: COLLECT_FRAMESCollect atomic positions or argument values from the trajectory for later analysis This action is a shortcut and uses the defaults shown here. More details STRIDE the frequency with which data should be stored for analysis=1 ATOMSlist of atomic positions that you would like to collect and store for later analysis=1-256  CLEAR the frequency with which data should all be deleted and restarted=0 ALIGN if storing atoms how would you like the alignment to be done can be SIMPLE/OPTIMAL=OPTIMAL
# PLUMED interprets the command:
# ff: COLLECT_FRAMES STRIDE=1 ATOMS=1-256
# as follows (Click the red comment above to revert to the short version of the input):
ff_getposxThe POSITION action with label ff_getposx calculates the following quantities:
 Quantity   
 Type   
 Description  
ff_getposx.x
vector
the x-component of the atom position
ff_getposx.y
vector
the y-component of the atom position
ff_getposx.z
vector
the z-component of the atom position: POSITIONCalculate the components of the position of an atom or atoms. More details ATOMSthe atom numbers that you would like to use the positions of=1-256
ff_getposThe CONCATENATE action with label ff_getpos calculates the following quantities:
 Quantity   
 Type   
 Description  
ff_getpos
vector
the concatenated vector/matrix that was constructed from the input values: CONCATENATEJoin vectors or matrices together More details ARGthe values that should be concatenated together to form the output vector=ff_getposx.x,ff_getposx.y,ff_getposx.z
ff_cposxThe MEAN action with label ff_cposx calculates the following quantities:
 Quantity   
 Type   
 Description  
ff_cposx
scalar
the MEAN of the elements in the input value: MEANCalculate the arithmetic mean of the elements in a vector More details ARGthe vector/matrix/grid whose elements shuld be added together=ff_getposx.x PERIODICif the output of your function is periodic then you should specify the periodicity of the function=NO
ff_cposyThe MEAN action with label ff_cposy calculates the following quantities:
 Quantity   
 Type   
 Description  
ff_cposy
scalar
the MEAN of the elements in the input value: MEANCalculate the arithmetic mean of the elements in a vector More details ARGthe vector/matrix/grid whose elements shuld be added together=ff_getposx.y PERIODICif the output of your function is periodic then you should specify the periodicity of the function=NO
ff_cposzThe MEAN action with label ff_cposz calculates the following quantities:
 Quantity   
 Type   
 Description  
ff_cposz
scalar
the MEAN of the elements in the input value: MEANCalculate the arithmetic mean of the elements in a vector More details ARGthe vector/matrix/grid whose elements shuld be added together=ff_getposx.z PERIODICif the output of your function is periodic then you should specify the periodicity of the function=NO
ff_refxThe CUSTOM action with label ff_refx calculates the following quantities:
 Quantity   
 Type   
 Description  
ff_refx
vector
the vector obtained by doing an element-wise application of an arbitrary function to the input vectors: CUSTOMCalculate a combination of variables using a custom expression. More details ARGthe values input to this function=ff_getposx.x,ff_cposx FUNCthe function you wish to evaluate=x-y PERIODICif the output of your function is periodic then you should specify the periodicity of the function=NO
ff_refyThe CUSTOM action with label ff_refy calculates the following quantities:
 Quantity   
 Type   
 Description  
ff_refy
vector
the vector obtained by doing an element-wise application of an arbitrary function to the input vectors: CUSTOMCalculate a combination of variables using a custom expression. More details ARGthe values input to this function=ff_getposx.y,ff_cposy FUNCthe function you wish to evaluate=x-y PERIODICif the output of your function is periodic then you should specify the periodicity of the function=NO
ff_refzThe CUSTOM action with label ff_refz calculates the following quantities:
 Quantity   
 Type   
 Description  
ff_refz
vector
the vector obtained by doing an element-wise application of an arbitrary function to the input vectors: CUSTOMCalculate a combination of variables using a custom expression. More details ARGthe values input to this function=ff_getposx.z,ff_cposz FUNCthe function you wish to evaluate=x-y PERIODICif the output of your function is periodic then you should specify the periodicity of the function=NO
ff_refThe CONCATENATE action with label ff_ref calculates the following quantities:
 Quantity   
 Type   
 Description  
ff_ref
vector
the concatenated vector/matrix that was constructed from the input values: CONCATENATEJoin vectors or matrices together More details ARGthe values that should be concatenated together to form the output vector=ff_refx,ff_refy,ff_refz
ff_refposThe COLLECT action with label ff_refpos calculates the following quantities:
 Quantity   
 Type   
 Description  
ff_refpos
matrix
the time series for the input quantity: COLLECTCollect data from the trajectory for later analysis More details TYPE required if you are collecting an object with rank>0=matrix ARGthe label of the value whose time series is being stored for later analysis=ff_ref STRIDE the frequency with which the data should be collected and added to the quantity being averaged=0 CLEAR the frequency with which to clear all the accumulated data=0
ff_refposTThe TRANSPOSE action with label ff_refposT calculates the following quantities:
 Quantity   
 Type   
 Description  
ff_refposT
vector
the transpose of the input matrix: TRANSPOSECalculate the transpose of a matrix More details ARGthe label of the vector or matrix that should be transposed=ff_refpos
ff_rmsdThe RMSD action with label ff_rmsd calculates the following quantities:
 Quantity   
 Type   
 Description  
ff_rmsd.dist
scalar
the RMSD distance the atoms have moved
ff_rmsd.disp
vector
the vector of displacements for the atoms: RMSD_VECTOR More details ARGthe labels of two actions that you are calculating the RMSD between=ff_getpos,ff_refpos DISPLACEMENT Calculate the vector of displacements instead of the length of this vector SQUARED  This should be set if you want mean squared displacement instead of RMSD  TYPE the manner in which RMSD alignment is performed=OPTIMAL
ff_fposThe COMBINE action with label ff_fpos calculates the following quantities:
 Quantity   
 Type   
 Description  
ff_fpos
vector
the vector obtained by doing an element-wise application of a linear combination to the input vectors: COMBINECalculate a polynomial combination of a set of other variables. More details ARGthe values input to this function=ff_refposT,ff_rmsd.disp PERIODICif the output of your function is periodic then you should specify the periodicity of the function=NO
ff_dataThe COLLECT action with label ff_data calculates the following quantities:
 Quantity   
 Type   
 Description  
ff_data
matrix
the time series for the input quantity: COLLECTCollect data from the trajectory for later analysis More details TYPE required if you are collecting an object with rank>0=matrix ARGthe label of the value whose time series is being stored for later analysis=ff_fpos STRIDE the frequency with which the data should be collected and added to the quantity being averaged=1 CLEAR the frequency with which to clear all the accumulated data=0
ff_cweightThe CONSTANT action with label ff_cweight calculates the following quantities:
 Quantity   
 Type   
 Description  
ff_cweight
scalar
the constant value that was read from the plumed input: CONSTANTCreate a constant value that can be passed to actions More details VALUEthe single number that you would like to store=0
ff_logweightsThe COLLECT action with label ff_logweights calculates the following quantities:
 Quantity   
 Type   
 Description  
ff_logweights
vector
the time series for the input quantity: COLLECTCollect data from the trajectory for later analysis More details ARGthe label of the value whose time series is being stored for later analysis=ff_cweight STRIDE the frequency with which the data should be collected and added to the quantity being averaged=1 CLEAR the frequency with which to clear all the accumulated data=0
ff_oneThe CONSTANT action with label ff_one calculates the following quantities:
 Quantity   
 Type   
 Description  
ff_one
scalar
the constant value that was read from the plumed input: CONSTANTCreate a constant value that can be passed to actions More details VALUEthe single number that you would like to store=1
ff_onesThe COLLECT action with label ff_ones calculates the following quantities:
 Quantity   
 Type   
 Description  
ff_ones
vector
the time series for the input quantity: COLLECTCollect data from the trajectory for later analysis More details ARGthe label of the value whose time series is being stored for later analysis=ff_one STRIDE the frequency with which the data should be collected and added to the quantity being averaged=1 CLEAR the frequency with which to clear all the accumulated data=0
# --- End of included input --- lweThe CUSTOM action with label lwe calculates the following quantities:
 Quantity   
 Type   
 Description  
lwe
vector
the vector obtained by doing an element-wise application of an arbitrary function to the input vectors: CUSTOMCalculate a combination of variables using a custom expression. More details ARGthe values input to this function=ff_logweights FUNCthe function you wish to evaluate=exp(x) PERIODICif the output of your function is periodic then you should specify the periodicity of the function=NO
smapThe SKETCHMAP action with label smap calculates the following quantities:
 Quantity   
 Type   
 Description  
smap
matrix
the sketch-map projection of the input points: SKETCHMAPConstruct a sketch map projection of the input data This action is a shortcut and it has hidden defaults. More details ...
   ARGthe matrix of high dimensional coordinates that you want to project in the low dimensional space=ff NLOW_DIMnumber of low-dimensional coordinates required=2
   HIGH_DIM_FUNCTIONthe parameters of the switching function in the high dimensional space={SMAP R_0=2 A=3 B=9}
   LOW_DIM_FUNCTIONthe parameters of the switching function in the low dimensional space={SMAP R_0=2 A=2 B=2}
   WEIGHTSa vector containing the weights of the points=lwe NCYCLES The number of cycles of pointwise global optimisation that are required=3 FGRID_SIZE interpolate the grid onto this number of points -- only works in 2D=50,50
...
smap: SKETCHMAPConstruct a sketch map projection of the input data This action is a shortcut and uses the defaults shown here. More details ...
   ARGthe matrix of high dimensional coordinates that you want to project in the low dimensional space=ff NLOW_DIMnumber of low-dimensional coordinates required=2
   HIGH_DIM_FUNCTIONthe parameters of the switching function in the high dimensional space={SMAP R_0=2 A=3 B=9}
   LOW_DIM_FUNCTIONthe parameters of the switching function in the low dimensional space={SMAP R_0=2 A=2 B=2}
   WEIGHTSa vector containing the weights of the points=lwe NCYCLES The number of cycles of pointwise global optimisation that are required=3 FGRID_SIZE interpolate the grid onto this number of points -- only works in 2D=50,50
 CGTOL The tolerance for the conjugate gradient minimization that finds the projection of the landmarks=1E-6 MAXITER maximum number of optimization cycles for optimisation algorithms=1000 BUFFER grid extent for search is (max projection - minimum projection) multiplied by this value=1.1 CGRID_SIZE number of points to use in each grid direction=10,10
...
# PLUMED interprets the command:
# smap: SKETCHMAP ...
#    ARG=ff NLOW_DIM=2
#    HIGH_DIM_FUNCTION={SMAP R_0=2 A=3 B=9}
#    LOW_DIM_FUNCTION={SMAP R_0=2 A=2 B=2}
#    WEIGHTS=lwe NCYCLES=3 FGRID_SIZE=50,50
# ...
# as follows (Click the red comment above to revert to the short version of the input):
smap_mdsThe VSTACK action with label smap_mds calculates the following quantities:
 Quantity   
 Type   
 Description  
smap_mds
matrix
a matrix that contains the input vectors in its columns: CLASSICAL_MDSCreate a low-dimensional projection of a trajectory using the classical multidimensional More details ARGthe arguments that you would like to make the histogram for=ff NLOW_DIMnumber of low-dimensional coordinates required=2
smap_hdmatThe MORE_THAN action with label smap_hdmat calculates the following quantities:
 Quantity   
 Type   
 Description  
smap_hdmat
matrix
the matrix obtained by doing an element-wise application of a function that is one if the if the input is more than a threshold to the input matrix: MORE_THANUse a switching function to determine how many of the input variables are more than a certain cutoff. More details ARGthe values input to this function=smap_mds_mat SQUARED is the input quantity the square of the value that you would like to apply the switching function to SWITCHThis keyword is used if you want to employ an alternative to the continuous swiching function defined above={SMAP R_0=2 A=3 B=9}
lwe_sumThe SUM action with label lwe_sum calculates the following quantities:
 Quantity   
 Type   
 Description  
lwe_sum
scalar
the SUM of the elements in the input value: SUMCalculate the sum of the arguments More details ARGthe vector/matrix/grid whose elements shuld be added together=lwe PERIODICif the output of your function is periodic then you should specify the periodicity of the function=NO
lwe_normedThe CUSTOM action with label lwe_normed calculates the following quantities:
 Quantity   
 Type   
 Description  
lwe_normed
vector
the vector obtained by doing an element-wise application of an arbitrary function to the input vectors: CUSTOMCalculate a combination of variables using a custom expression. More details ARGthe values input to this function=lwe,lwe_sum FUNCthe function you wish to evaluate=x/y PERIODICif the output of your function is periodic then you should specify the periodicity of the function=NO
lwe_matThe OUTER_PRODUCT action with label lwe_mat calculates the following quantities:
 Quantity   
 Type   
 Description  
lwe_mat
matrix
a matrix containing the outer product of the two input vectors that was obtained using the function that was input: OUTER_PRODUCTCalculate the outer product matrix of two vectors More details ARGthe labels of the two vectors from which the outer product is being computed=lwe_normed,lwe_normed
smap_cgThe ARRANGE_POINTS action with label smap_cg calculates the following quantities:
 Quantity   
 Type   
 Description  
smap_cg.coord-1
vector
the coordinates of the points in the low dimensional space  This is the 1th of these quantities
smap_cg.coord-2
vector
the coordinates of the points in the low dimensional space  This is the 2th of these quantities: ARRANGE_POINTSArrange points in a low dimensional space so that the (transformed) distances between points in the low dimensional space match the dissimilarities provided in an input matrix. More details ARGthe initial positions for the projections=smap_mds-1,smap_mds-2 MINTYPE the method to use for the minimisation=conjgrad TARGET1the matrix of target quantities that you would like to match=smap_hdmat FUNC1a function that is applied on the distances between the points in the low dimensional space={SMAP R_0=2 A=2 B=2} WEIGHTS1the matrix with the weights of the target quantities=lwe_mat CGTOL the tolerance for the conjugate gradient minimization=1E-6
smap_apThe ARRANGE_POINTS action with label smap_ap calculates the following quantities:
 Quantity   
 Type   
 Description  
smap_ap.coord-1
vector
the coordinates of the points in the low dimensional space  This is the 1th of these quantities
smap_ap.coord-2
vector
the coordinates of the points in the low dimensional space  This is the 2th of these quantities: ARRANGE_POINTSArrange points in a low dimensional space so that the (transformed) distances between points in the low dimensional space match the dissimilarities provided in an input matrix. More details ARGthe initial positions for the projections=smap_cg.coord-1,smap_cg.coord-2 NCYCLES the number of cycles of global optimization to attempt=3 BUFFER grid extent for search is (max projection - minimum projection) multiplied by this value=1.1 FGRID_SIZE interpolate the grid onto this number of points -- only works in 2D=50,50 CGRID_SIZE number of points to use in each grid direction=10,10 MINTYPE the method to use for the minimisation=pointwise TARGET1the matrix of target quantities that you would like to match=smap_hdmat FUNC1a function that is applied on the distances between the points in the low dimensional space={SMAP R_0=2 A=2 B=2} WEIGHTS1the matrix with the weights of the target quantities=lwe_mat CGTOL the tolerance for the conjugate gradient minimization=1E-6
smapThe VSTACK action with label smap calculates the following quantities:
 Quantity   
 Type   
 Description  
smap
matrix
a matrix that contains the input vectors in its columns: VSTACKCreate a matrix by stacking vectors together More details ARGthe values that you would like to stack together to construct the output matrix=smap_ap.coord-1,smap_ap.coord-2
# --- End of included input --- 
DUMPVECTORPrint a vector to a file More details ARGthe labels of vectors/matrices that should be output in the file=smap.*,lwe FILE the file on which to write the vetors=smap

Information on how we optimise the sketch-map stress function can be found by expanding the SKETCHMAP shortcut in the above input and reading the documentation for ARRANGE_POINTS.

Using landmarks

Optimising the sketch-map stress function is computationally expensive and so the usual practise with this method is to pick a subset of landmark points. Projections for these points are found by optimising the sketch-map stress function using ARRANGE_POINTS. Projections for non-landmark points are then found by using PROJECT_POINTS. The following example input illustrates how you can perform such a calculation with PLUMED

Click on the labels of the actions for more information on what each action computes

d1The DISTANCE action with label d1 calculates the following quantities:
 Quantity   
 Type   
 Description  
d1
scalar
the DISTANCE between this pair of atoms: DISTANCECalculate the distance/s between pairs of atoms. More details ATOMSthe pair of atom that we are calculating the distance between=1,2
d2The DISTANCE action with label d2 calculates the following quantities:
 Quantity   
 Type   
 Description  
d2
scalar
the DISTANCE between this pair of atoms: DISTANCECalculate the distance/s between pairs of atoms. More details ATOMSthe pair of atom that we are calculating the distance between=3,4
d3The DISTANCE action with label d3 calculates the following quantities:
 Quantity   
 Type   
 Description  
d3
scalar
the DISTANCE between this pair of atoms: DISTANCECalculate the distance/s between pairs of atoms. More details ATOMSthe pair of atom that we are calculating the distance between=5,6

ffThe COLLECT_FRAMES action with label ff calculates the following quantities:
 Quantity   
 Type   
 Description  
ff_data
matrix
the data that is being collected by this action
ff_logweights
vector
the logarithms of the weights of the data points: COLLECT_FRAMESCollect atomic positions or argument values from the trajectory for later analysis This action is a shortcut and it has hidden defaults. More details ARGthe labels of the values whose time series you would like to collect for later analysis=d1,d2,d3
The COLLECT_FRAMES action with label ff calculates the following quantities:
 Quantity   
 Description  
ff.data
the data that is being collected by this action
ff.logweights
the logarithms of the weights of the data pointsff: COLLECT_FRAMESCollect atomic positions or argument values from the trajectory for later analysis This action is a shortcut and uses the defaults shown here. More details ARGthe labels of the values whose time series you would like to collect for later analysis=d1,d2,d3  STRIDE the frequency with which data should be stored for analysis=1 CLEAR the frequency with which data should all be deleted and restarted=0
# PLUMED interprets the command:
# ff: COLLECT_FRAMES ARG=d1,d2,d3
# as follows (Click the red comment above to revert to the short version of the input):
ff_d1The COLLECT action with label ff_d1 calculates the following quantities:
 Quantity   
 Type   
 Description  
ff_d1
vector
the time series for the input quantity: COLLECTCollect data from the trajectory for later analysis More details ARGthe label of the value whose time series is being stored for later analysis=d1 STRIDE the frequency with which the data should be collected and added to the quantity being averaged=1 CLEAR the frequency with which to clear all the accumulated data=0
ff_d2The COLLECT action with label ff_d2 calculates the following quantities:
 Quantity   
 Type   
 Description  
ff_d2
vector
the time series for the input quantity: COLLECTCollect data from the trajectory for later analysis More details ARGthe label of the value whose time series is being stored for later analysis=d2 STRIDE the frequency with which the data should be collected and added to the quantity being averaged=1 CLEAR the frequency with which to clear all the accumulated data=0
ff_d3The COLLECT action with label ff_d3 calculates the following quantities:
 Quantity   
 Type   
 Description  
ff_d3
vector
the time series for the input quantity: COLLECTCollect data from the trajectory for later analysis More details ARGthe label of the value whose time series is being stored for later analysis=d3 STRIDE the frequency with which the data should be collected and added to the quantity being averaged=1 CLEAR the frequency with which to clear all the accumulated data=0
ff_dataThe VSTACK action with label ff_data calculates the following quantities:
 Quantity   
 Type   
 Description  
ff_data
matrix
a matrix that contains the input vectors in its columns: VSTACKCreate a matrix by stacking vectors together More details ARGthe values that you would like to stack together to construct the output matrix=ff_d1,ff_d2,ff_d3
ff_cweightThe CONSTANT action with label ff_cweight calculates the following quantities:
 Quantity   
 Type   
 Description  
ff_cweight
scalar
the constant value that was read from the plumed input: CONSTANTCreate a constant value that can be passed to actions More details VALUEthe single number that you would like to store=0
ff_logweightsThe COLLECT action with label ff_logweights calculates the following quantities:
 Quantity   
 Type   
 Description  
ff_logweights
vector
the time series for the input quantity: COLLECTCollect data from the trajectory for later analysis More details ARGthe label of the value whose time series is being stored for later analysis=ff_cweight STRIDE the frequency with which the data should be collected and added to the quantity being averaged=1 CLEAR the frequency with which to clear all the accumulated data=0
ff_oneThe CONSTANT action with label ff_one calculates the following quantities:
 Quantity   
 Type   
 Description  
ff_one
scalar
the constant value that was read from the plumed input: CONSTANTCreate a constant value that can be passed to actions More details VALUEthe single number that you would like to store=1
ff_onesThe COLLECT action with label ff_ones calculates the following quantities:
 Quantity   
 Type   
 Description  
ff_ones
vector
the time series for the input quantity: COLLECTCollect data from the trajectory for later analysis More details ARGthe label of the value whose time series is being stored for later analysis=ff_one STRIDE the frequency with which the data should be collected and added to the quantity being averaged=1 CLEAR the frequency with which to clear all the accumulated data=0
# --- End of included input --- ff_dataTThe TRANSPOSE action with label ff_dataT calculates the following quantities:
 Quantity   
 Type   
 Description  
ff_dataT
matrix
the transpose of the input matrix: TRANSPOSECalculate the transpose of a matrix More details ARGthe label of the vector or matrix that should be transposed=ff_data
llThe LANDMARK_SELECT_STRIDE action with label ll calculates the following quantities:
 Quantity   
 Type   
 Description  
ll_data
matrix
the data that is being collected by this action
ll_rectdissims
matrix
a rectangular matrix containing the distances between the landmark points and the rest of the points
ll_sqrdissims
matrix
a square matrix containing the distances between each pair of landmark points
ll_logweights
vector
the logarithms of the weights of the data points: LANDMARK_SELECT_STRIDESelect every ith frame from the stored set of configurations This action is a shortcut. More details ARGthe COLLECT_FRAMES action that you used to get the data=ff NLANDMARKSthe numbe rof landmarks you would like to create=250
# PLUMED interprets the command:
# ll: LANDMARK_SELECT_STRIDE ARG=ff NLANDMARKS=250
# as follows (Click the red comment above to revert to the short version of the input):
The LANDMARK_SELECT_STRIDE action with label ll calculates the following quantities:
 Quantity   
 Description  
ll.data
the data that is being collected by this action
ll.logweights
the logarithms of the weights of the data pointsll_allweightsThe CUSTOM action with label ll_allweights calculates the following quantities:
 Quantity   
 Type   
 Description  
ll_allweights
vector
the vector obtained by doing an element-wise application of an arbitrary function to the input vectors: LOGSUMEXPThis action takes the exponential of a vector of logarithms and divides each element of the vector by the sum of the exponentials. More details ARGthe vector of logweights that you would like to normalise using the logsumexp trick=ff_logweights
ll_ff_dataTThe TRANSPOSE action with label ll_ff_dataT calculates the following quantities:
 Quantity   
 Type   
 Description  
ll_ff_dataT
matrix
the transpose of the input matrix: TRANSPOSECalculate the transpose of a matrix More details ARGthe label of the vector or matrix that should be transposed=ff_data
ll_dissimsThe DISSIMILARITIES action with label ll_dissims calculates the following quantities:
 Quantity   
 Type   
 Description  
ll_dissims
matrix
the dissimilarity matrix: DISSIMILARITIESCalculate the matrix of dissimilarities between a trajectory of atomic configurations. More details SQUARED calculate the squares of the dissimilarities (this option cannot be used with MATRIX_PRODUCT) ARGthe label of the two matrices from which the product is calculated=ff_data,ll_ff_dataT
ll_maskThe CREATE_MASK action with label ll_mask calculates the following quantities:
 Quantity   
 Type   
 Description  
ll_mask
vector
a vector of zeros and ones that is used that can be used to mask some of the elements in a time series: CREATE_MASKCreate a mask vector to use for landmark selection More details ARGthe label of the vector that you would like to construct a mask for=ll_allweights TYPEthe way the zeros are supposed to be set=stride NZEROSthe number of zeros that you want to put in the mask=250
ll_dataThe SELECT_WITH_MASK action with label ll_data calculates the following quantities:
 Quantity   
 Type   
 Description  
ll_data
matrix
a vector/matrix of values that is obtained using a mask to select elements of interest: SELECT_WITH_MASKUse a mask to select elements of an array More details ARGthe label for the value upon which you are going to apply the mask=ff_data ROW_MASKan array with ones in the rows of the matrix that you want to discard=ll_mask
ll_rmaskThe CREATE_MASK action with label ll_rmask calculates the following quantities:
 Quantity   
 Type   
 Description  
ll_rmask
vector
a vector of zeros and ones that is used that can be used to mask some of the elements in a time series: CREATE_MASKCreate a mask vector to use for landmark selection More details ARGthe label of the vector that you would like to construct a mask for=ll_allweights TYPEthe way the zeros are supposed to be set=nomask
ll_rectdissimsThe SELECT_WITH_MASK action with label ll_rectdissims calculates the following quantities:
 Quantity   
 Type   
 Description  
ll_rectdissims
matrix
a vector/matrix of values that is obtained using a mask to select elements of interest: SELECT_WITH_MASKUse a mask to select elements of an array More details ARGthe label for the value upon which you are going to apply the mask=ll_dissims COLUMN_MASKan array with ones in the columns of the matrix that you want to discard=ll_mask ROW_MASKan array with ones in the rows of the matrix that you want to discard=ll_rmask
ll_sqrdissimsThe SELECT_WITH_MASK action with label ll_sqrdissims calculates the following quantities:
 Quantity   
 Type   
 Description  
ll_sqrdissims
matrix
a vector/matrix of values that is obtained using a mask to select elements of interest: SELECT_WITH_MASKUse a mask to select elements of an array More details ARGthe label for the value upon which you are going to apply the mask=ll_dissims ROW_MASKan array with ones in the rows of the matrix that you want to discard=ll_mask COLUMN_MASKan array with ones in the columns of the matrix that you want to discard=ll_mask
ll_voronoiThe NEIGHBORS action with label ll_voronoi calculates the following quantities:
 Quantity   
 Type   
 Description  
ll_voronoi
matrix
a matrix in which the ij element is one if the ij-element of the input matrix is one of the NLOWEST/NHIGHEST elements on that row of the input matrix and zero otherwise: VORONOIDo a voronoi analysis More details ARGthe distance/adjacency matrix that should be used to perform the voronoi analysis=ll_rectdissims
ll_allweightsTThe TRANSPOSE action with label ll_allweightsT calculates the following quantities:
 Quantity   
 Type   
 Description  
ll_allweightsT
matrix
the transpose of the input matrix: TRANSPOSECalculate the transpose of a matrix More details ARGthe label of the vector or matrix that should be transposed=ll_allweights
ll_weightsTThe MATRIX_PRODUCT action with label ll_weightsT calculates the following quantities:
 Quantity   
 Type   
 Description  
ll_weightsT
matrix
the product of the two input matrices: MATRIX_PRODUCTCalculate the product of two matrices More details ARGthe label of the two matrices from which the product is calculated=ll_allweightsT,ll_voronoi
ll_weightsThe TRANSPOSE action with label ll_weights calculates the following quantities:
 Quantity   
 Type   
 Description  
ll_weights
vector
the transpose of the input matrix: TRANSPOSECalculate the transpose of a matrix More details ARGthe label of the vector or matrix that should be transposed=ll_weightsT
ll_logweightsThe CUSTOM action with label ll_logweights calculates the following quantities:
 Quantity   
 Type   
 Description  
ll_logweights
vector
the vector obtained by doing an element-wise application of an arbitrary function to the input vectors: CUSTOMCalculate a combination of variables using a custom expression. More details ARGthe values input to this function=ll_weights FUNCthe function you wish to evaluate=log(x) PERIODICif the output of your function is periodic then you should specify the periodicity of the function=NO
ll_onesThe SELECT_WITH_MASK action with label ll_ones calculates the following quantities:
 Quantity   
 Type   
 Description  
ll_ones
vector
a vector/matrix of values that is obtained using a mask to select elements of interest: SELECT_WITH_MASKUse a mask to select elements of an array More details ARGthe label for the value upon which you are going to apply the mask=ff_ones MASKan array with ones in the components that you want to discard=ll_mask
# --- End of included input --- 
# Calculate the weights
voroThe VORONOI action with label voro calculates the following quantities:
 Quantity   
 Type   
 Description  
voro
matrix
a matrix in which element ij is equal to one if the ij component of the input matrix is lower than all the ik elements of the matrix where k is not j and zero otherwise: VORONOIDo a voronoi analysis This action is a shortcut. More details ARGthe distance/adjacency matrix that should be used to perform the voronoi analysis=ll_rectdissims
# PLUMED interprets the command:
# voro: VORONOI ARG=ll_rectdissims
# as follows (Click the red comment above to revert to the short version of the input):
voroThe NEIGHBORS action with label voro calculates the following quantities:
 Quantity   
 Type   
 Description  
voro
matrix
a matrix in which the ij element is one if the ij-element of the input matrix is one of the NLOWEST/NHIGHEST elements on that row of the input matrix and zero otherwise: NEIGHBORSBuild a matrix with ones in for the N nearest neighbours of an atom More details NLOWEST in each row of the output matrix set the elements that correspond to the n lowest elements in each row of the input matrix equal to one=1  ARGthe label of an adjacency/distance matrix that will be used to find the nearest neighbors=ll_rectdissims
# --- End of included input --- weightsThe CUSTOM action with label weights calculates the following quantities:
 Quantity   
 Type   
 Description  
weights
vector
the vector obtained by doing an element-wise application of an arbitrary function to the input vectors: CUSTOMCalculate a combination of variables using a custom expression. More details ARGthe values input to this function=ff.logweights FUNCthe function you wish to evaluate=exp(x) PERIODICif the output of your function is periodic then you should specify the periodicity of the function=NO
weightsTThe TRANSPOSE action with label weightsT calculates the following quantities:
 Quantity   
 Type   
 Description  
weightsT
matrix
the transpose of the input matrix: TRANSPOSECalculate the transpose of a matrix More details ARGthe label of the vector or matrix that should be transposed=weights
lweTThe MATRIX_PRODUCT action with label lweT calculates the following quantities:
 Quantity   
 Type   
 Description  
lweT
matrix
the product of the two input matrices: MATRIX_PRODUCTCalculate the product of two matrices More details ARGthe label of the two matrices from which the product is calculated=weightsT,voro
lweThe TRANSPOSE action with label lwe calculates the following quantities:
 Quantity   
 Type   
 Description  
lwe
vector
the transpose of the input matrix: TRANSPOSECalculate the transpose of a matrix More details ARGthe label of the vector or matrix that should be transposed=lweT

smapThe SKETCHMAP action with label smap calculates the following quantities:
 Quantity   
 Type   
 Description  
smap
matrix
the sketch-map projection of the input points
smap_osample
matrix
the out-of-sample projections: SKETCHMAPConstruct a sketch map projection of the input data This action is a shortcut and it has hidden defaults. More details ...
  ARGthe matrix of high dimensional coordinates that you want to project in the low dimensional space=ll NLOW_DIMnumber of low-dimensional coordinates required=2 PROJECT_ALL if the input are landmark coordinates then project the out of sample configurations
  HIGH_DIM_FUNCTIONthe parameters of the switching function in the high dimensional space={SMAP R_0=4 A=3 B=2}
  LOW_DIM_FUNCTIONthe parameters of the switching function in the low dimensional space={SMAP R_0=4 A=1 B=2}
...
smap: SKETCHMAPConstruct a sketch map projection of the input data This action is a shortcut and uses the defaults shown here. More details ...
  ARGthe matrix of high dimensional coordinates that you want to project in the low dimensional space=ll NLOW_DIMnumber of low-dimensional coordinates required=2 PROJECT_ALL if the input are landmark coordinates then project the out of sample configurations
  HIGH_DIM_FUNCTIONthe parameters of the switching function in the high dimensional space={SMAP R_0=4 A=3 B=2}
  LOW_DIM_FUNCTIONthe parameters of the switching function in the low dimensional space={SMAP R_0=4 A=1 B=2}
 CGTOL The tolerance for the conjugate gradient minimization that finds the projection of the landmarks=1E-6 MAXITER maximum number of optimization cycles for optimisation algorithms=1000 NCYCLES The number of cycles of pointwise global optimisation that are required=0 OS_CGTOL The tolerance for the conjugate gradient minimization that finds the out of sample projections=1E-6
...
# PLUMED interprets the command:
# smap: SKETCHMAP ...
#   ARG=ll NLOW_DIM=2 PROJECT_ALL
#   HIGH_DIM_FUNCTION={SMAP R_0=4 A=3 B=2}
#   LOW_DIM_FUNCTION={SMAP R_0=4 A=1 B=2}
# ...
# as follows (Click the red comment above to revert to the short version of the input):
smap_mdsThe VSTACK action with label smap_mds calculates the following quantities:
 Quantity   
 Type   
 Description  
smap_mds
matrix
a matrix that contains the input vectors in its columns: CLASSICAL_MDSCreate a low-dimensional projection of a trajectory using the classical multidimensional More details ARGthe arguments that you would like to make the histogram for=ll NLOW_DIMnumber of low-dimensional coordinates required=2
smap_hdmatThe MORE_THAN action with label smap_hdmat calculates the following quantities:
 Quantity   
 Type   
 Description  
smap_hdmat
matrix
the matrix obtained by doing an element-wise application of a function that is one if the if the input is more than a threshold to the input matrix: MORE_THANUse a switching function to determine how many of the input variables are more than a certain cutoff. More details ARGthe values input to this function=ll_sqrdissims SQUARED is the input quantity the square of the value that you would like to apply the switching function to SWITCHThis keyword is used if you want to employ an alternative to the continuous swiching function defined above={SMAP R_0=4 A=3 B=2}
ll_weights_sumThe SUM action with label ll_weights_sum calculates the following quantities:
 Quantity   
 Type   
 Description  
ll_weights_sum
scalar
the SUM of the elements in the input value: SUMCalculate the sum of the arguments More details ARGthe vector/matrix/grid whose elements shuld be added together=ll_weights PERIODICif the output of your function is periodic then you should specify the periodicity of the function=NO
ll_weights_normedThe CUSTOM action with label ll_weights_normed calculates the following quantities:
 Quantity   
 Type   
 Description  
ll_weights_normed
vector
the vector obtained by doing an element-wise application of an arbitrary function to the input vectors: CUSTOMCalculate a combination of variables using a custom expression. More details ARGthe values input to this function=ll_weights,ll_weights_sum FUNCthe function you wish to evaluate=x/y PERIODICif the output of your function is periodic then you should specify the periodicity of the function=NO
ll_weights_matThe OUTER_PRODUCT action with label ll_weights_mat calculates the following quantities:
 Quantity   
 Type   
 Description  
ll_weights_mat
matrix
a matrix containing the outer product of the two input vectors that was obtained using the function that was input: OUTER_PRODUCTCalculate the outer product matrix of two vectors More details ARGthe labels of the two vectors from which the outer product is being computed=ll_weights_normed,ll_weights_normed
smap_apThe ARRANGE_POINTS action with label smap_ap calculates the following quantities:
 Quantity   
 Type   
 Description  
smap_ap.coord-1
vector
the coordinates of the points in the low dimensional space  This is the 1th of these quantities
smap_ap.coord-2
vector
the coordinates of the points in the low dimensional space  This is the 2th of these quantities: ARRANGE_POINTSArrange points in a low dimensional space so that the (transformed) distances between points in the low dimensional space match the dissimilarities provided in an input matrix. More details ARGthe initial positions for the projections=smap_mds-1,smap_mds-2 MINTYPE the method to use for the minimisation=conjgrad TARGET1the matrix of target quantities that you would like to match=smap_hdmat FUNC1a function that is applied on the distances between the points in the low dimensional space={SMAP R_0=4 A=1 B=2} WEIGHTS1the matrix with the weights of the target quantities=ll_weights_mat CGTOL the tolerance for the conjugate gradient minimization=1E-6
smapThe VSTACK action with label smap calculates the following quantities:
 Quantity   
 Type   
 Description  
smap
matrix
a matrix that contains the input vectors in its columns: VSTACKCreate a matrix by stacking vectors together More details ARGthe values that you would like to stack together to construct the output matrix=smap_ap.coord-1,smap_ap.coord-2
smap_lhdmatThe MORE_THAN action with label smap_lhdmat calculates the following quantities:
 Quantity   
 Type   
 Description  
smap_lhdmat
matrix
the matrix obtained by doing an element-wise application of a function that is one if the if the input is more than a threshold to the input matrix: MORE_THANUse a switching function to determine how many of the input variables are more than a certain cutoff. More details ARGthe values input to this function=ll_rectdissims SQUARED is the input quantity the square of the value that you would like to apply the switching function to SWITCHThis keyword is used if you want to employ an alternative to the continuous swiching function defined above={SMAP R_0=4 A=3 B=2}
smap_osample_ppThe PROJECT_POINTS action with label smap_osample_pp calculates the following quantities:
 Quantity   
 Type   
 Description  
smap_osample_pp.coord-1
vector
the coordinates of the points in the low dimensional space  This is the 1th of these quantities
smap_osample_pp.coord-2
vector
the coordinates of the points in the low dimensional space  This is the 2th of these quantities: PROJECT_POINTSFind the projection of a point in a low dimensional space by matching the (transformed) distance between it and a series of reference configurations that were input More details ARGthe projections of the landmark points=smap_ap.coord-1,smap_ap.coord-2 TARGET1the matrix of target quantities that you would like to match=smap_lhdmat FUNC1a function that is applied on the distances between the points in the low dimensional space={SMAP R_0=4 A=1 B=2} WEIGHTS1the matrix with the weights of the target quantities=ll_weights_normed CGTOL the tolerance for the conjugate gradient minimization=1E-6
smap_osampleThe VSTACK action with label smap_osample calculates the following quantities:
 Quantity   
 Type   
 Description  
smap_osample
matrix
a matrix that contains the input vectors in its columns: VSTACKCreate a matrix by stacking vectors together More details ARGthe values that you would like to stack together to construct the output matrix=smap_osample_pp.coord-1,smap_osample_pp.coord-2
# --- End of included input --- 
DUMPVECTORPrint a vector to a file More details ARGthe labels of vectors/matrices that should be output in the file=smap,lwe FILE the file on which to write the vetors=smap
The DUMPVECTOR action with label  calculates somethingDUMPVECTORPrint a vector to a file More details ARGthe labels of vectors/matrices that should be output in the file=smap_osample,weights FILE the file on which to write the vetors=projections

Using SMACOF

By default we PLUMED uses the method described in the paper cited below to optimise the sketch-map stress function. In other words, we use a combination of conjugate gradients and a pointwise global optimisation algorithm. Within the code there is also an experimental implementation of optimisation using a variant on the smacof algorithm. If you would like to experiment with this option you use the USE_SMACOF flag as illustrated below:

Click on the labels of the actions for more information on what each action computes

d1The DISTANCE action with label d1 calculates the following quantities:
 Quantity   
 Type   
 Description  
d1
scalar
the DISTANCE between this pair of atoms: DISTANCECalculate the distance/s between pairs of atoms. More details ATOMSthe pair of atom that we are calculating the distance between=1,2
d2The DISTANCE action with label d2 calculates the following quantities:
 Quantity   
 Type   
 Description  
d2
scalar
the DISTANCE between this pair of atoms: DISTANCECalculate the distance/s between pairs of atoms. More details ATOMSthe pair of atom that we are calculating the distance between=3,4
d3The DISTANCE action with label d3 calculates the following quantities:
 Quantity   
 Type   
 Description  
d3
scalar
the DISTANCE between this pair of atoms: DISTANCECalculate the distance/s between pairs of atoms. More details ATOMSthe pair of atom that we are calculating the distance between=5,6

ffThe COLLECT_FRAMES action with label ff calculates the following quantities:
 Quantity   
 Type   
 Description  
ff_data
matrix
the data that is being collected by this action
ff_logweights
vector
the logarithms of the weights of the data points: COLLECT_FRAMESCollect atomic positions or argument values from the trajectory for later analysis This action is a shortcut and it has hidden defaults. More details STRIDE the frequency with which data should be stored for analysis=1 ARGthe labels of the values whose time series you would like to collect for later analysis=d1,d2,d3
The COLLECT_FRAMES action with label ff calculates the following quantities:
 Quantity   
 Description  
ff.data
the data that is being collected by this action
ff.logweights
the logarithms of the weights of the data pointsff: COLLECT_FRAMESCollect atomic positions or argument values from the trajectory for later analysis This action is a shortcut and uses the defaults shown here. More details STRIDE the frequency with which data should be stored for analysis=1 ARGthe labels of the values whose time series you would like to collect for later analysis=d1,d2,d3  CLEAR the frequency with which data should all be deleted and restarted=0
# PLUMED interprets the command:
# ff: COLLECT_FRAMES STRIDE=1 ARG=d1,d2,d3
# as follows (Click the red comment above to revert to the short version of the input):
ff_d1The COLLECT action with label ff_d1 calculates the following quantities:
 Quantity   
 Type   
 Description  
ff_d1
vector
the time series for the input quantity: COLLECTCollect data from the trajectory for later analysis More details ARGthe label of the value whose time series is being stored for later analysis=d1 STRIDE the frequency with which the data should be collected and added to the quantity being averaged=1 CLEAR the frequency with which to clear all the accumulated data=0
ff_d2The COLLECT action with label ff_d2 calculates the following quantities:
 Quantity   
 Type   
 Description  
ff_d2
vector
the time series for the input quantity: COLLECTCollect data from the trajectory for later analysis More details ARGthe label of the value whose time series is being stored for later analysis=d2 STRIDE the frequency with which the data should be collected and added to the quantity being averaged=1 CLEAR the frequency with which to clear all the accumulated data=0
ff_d3The COLLECT action with label ff_d3 calculates the following quantities:
 Quantity   
 Type   
 Description  
ff_d3
vector
the time series for the input quantity: COLLECTCollect data from the trajectory for later analysis More details ARGthe label of the value whose time series is being stored for later analysis=d3 STRIDE the frequency with which the data should be collected and added to the quantity being averaged=1 CLEAR the frequency with which to clear all the accumulated data=0
ff_dataThe VSTACK action with label ff_data calculates the following quantities:
 Quantity   
 Type   
 Description  
ff_data
matrix
a matrix that contains the input vectors in its columns: VSTACKCreate a matrix by stacking vectors together More details ARGthe values that you would like to stack together to construct the output matrix=ff_d1,ff_d2,ff_d3
ff_cweightThe CONSTANT action with label ff_cweight calculates the following quantities:
 Quantity   
 Type   
 Description  
ff_cweight
scalar
the constant value that was read from the plumed input: CONSTANTCreate a constant value that can be passed to actions More details VALUEthe single number that you would like to store=0
ff_logweightsThe COLLECT action with label ff_logweights calculates the following quantities:
 Quantity   
 Type   
 Description  
ff_logweights
vector
the time series for the input quantity: COLLECTCollect data from the trajectory for later analysis More details ARGthe label of the value whose time series is being stored for later analysis=ff_cweight STRIDE the frequency with which the data should be collected and added to the quantity being averaged=1 CLEAR the frequency with which to clear all the accumulated data=0
ff_oneThe CONSTANT action with label ff_one calculates the following quantities:
 Quantity   
 Type   
 Description  
ff_one
scalar
the constant value that was read from the plumed input: CONSTANTCreate a constant value that can be passed to actions More details VALUEthe single number that you would like to store=1
ff_onesThe COLLECT action with label ff_ones calculates the following quantities:
 Quantity   
 Type   
 Description  
ff_ones
vector
the time series for the input quantity: COLLECTCollect data from the trajectory for later analysis More details ARGthe label of the value whose time series is being stored for later analysis=ff_one STRIDE the frequency with which the data should be collected and added to the quantity being averaged=1 CLEAR the frequency with which to clear all the accumulated data=0
# --- End of included input --- lweThe CUSTOM action with label lwe calculates the following quantities:
 Quantity   
 Type   
 Description  
lwe
vector
the vector obtained by doing an element-wise application of an arbitrary function to the input vectors: CUSTOMCalculate a combination of variables using a custom expression. More details ARGthe values input to this function=ff_logweights FUNCthe function you wish to evaluate=exp(x) PERIODICif the output of your function is periodic then you should specify the periodicity of the function=NO
smapThe SKETCHMAP action with label smap calculates the following quantities:
 Quantity   
 Type   
 Description  
smap
matrix
the sketch-map projection of the input points: SKETCHMAPConstruct a sketch map projection of the input data This action is a shortcut and it has hidden defaults. More details ...
   ARGthe matrix of high dimensional coordinates that you want to project in the low dimensional space=ff NLOW_DIMnumber of low-dimensional coordinates required=2 USE_SMACOF find the projection in the low dimensional space using the SMACOF algorithm
   HIGH_DIM_FUNCTIONthe parameters of the switching function in the high dimensional space={SMAP R_0=2 A=3 B=9}
   LOW_DIM_FUNCTIONthe parameters of the switching function in the low dimensional space={SMAP R_0=2 A=2 B=2}
   WEIGHTSa vector containing the weights of the points=lwe
...
smap: SKETCHMAPConstruct a sketch map projection of the input data This action is a shortcut and uses the defaults shown here. More details ...
   ARGthe matrix of high dimensional coordinates that you want to project in the low dimensional space=ff NLOW_DIMnumber of low-dimensional coordinates required=2 USE_SMACOF find the projection in the low dimensional space using the SMACOF algorithm
   HIGH_DIM_FUNCTIONthe parameters of the switching function in the high dimensional space={SMAP R_0=2 A=3 B=9}
   LOW_DIM_FUNCTIONthe parameters of the switching function in the low dimensional space={SMAP R_0=2 A=2 B=2}
   WEIGHTSa vector containing the weights of the points=lwe
 CGTOL The tolerance for the conjugate gradient minimization that finds the projection of the landmarks=1E-6 MAXITER maximum number of optimization cycles for optimisation algorithms=1000 NCYCLES The number of cycles of pointwise global optimisation that are required=0 SMACTOL the tolerance for the smacof algorithm=1E-4 SMACREG this is used to ensure that we don't divide by zero when updating weights for SMACOF algorithm=0.001
...
# PLUMED interprets the command:
# smap: SKETCHMAP ...
#    ARG=ff NLOW_DIM=2 USE_SMACOF
#    HIGH_DIM_FUNCTION={SMAP R_0=2 A=3 B=9}
#    LOW_DIM_FUNCTION={SMAP R_0=2 A=2 B=2}
#    WEIGHTS=lwe
# ...
# as follows (Click the red comment above to revert to the short version of the input):
smap_mdsThe VSTACK action with label smap_mds calculates the following quantities:
 Quantity   
 Type   
 Description  
smap_mds
matrix
a matrix that contains the input vectors in its columns: CLASSICAL_MDSCreate a low-dimensional projection of a trajectory using the classical multidimensional More details ARGthe arguments that you would like to make the histogram for=ff NLOW_DIMnumber of low-dimensional coordinates required=2
smap_hdmatThe MORE_THAN action with label smap_hdmat calculates the following quantities:
 Quantity   
 Type   
 Description  
smap_hdmat
matrix
the matrix obtained by doing an element-wise application of a function that is one if the if the input is more than a threshold to the input matrix: MORE_THANUse a switching function to determine how many of the input variables are more than a certain cutoff. More details ARGthe values input to this function=smap_mds_mat SQUARED is the input quantity the square of the value that you would like to apply the switching function to SWITCHThis keyword is used if you want to employ an alternative to the continuous swiching function defined above={SMAP R_0=2 A=3 B=9}
lwe_sumThe SUM action with label lwe_sum calculates the following quantities:
 Quantity   
 Type   
 Description  
lwe_sum
scalar
the SUM of the elements in the input value: SUMCalculate the sum of the arguments More details ARGthe vector/matrix/grid whose elements shuld be added together=lwe PERIODICif the output of your function is periodic then you should specify the periodicity of the function=NO
lwe_normedThe CUSTOM action with label lwe_normed calculates the following quantities:
 Quantity   
 Type   
 Description  
lwe_normed
vector
the vector obtained by doing an element-wise application of an arbitrary function to the input vectors: CUSTOMCalculate a combination of variables using a custom expression. More details ARGthe values input to this function=lwe,lwe_sum FUNCthe function you wish to evaluate=x/y PERIODICif the output of your function is periodic then you should specify the periodicity of the function=NO
lwe_matThe OUTER_PRODUCT action with label lwe_mat calculates the following quantities:
 Quantity   
 Type   
 Description  
lwe_mat
matrix
a matrix containing the outer product of the two input vectors that was obtained using the function that was input: OUTER_PRODUCTCalculate the outer product matrix of two vectors More details ARGthe labels of the two vectors from which the outer product is being computed=lwe_normed,lwe_normed
smap_apThe ARRANGE_POINTS action with label smap_ap calculates the following quantities:
 Quantity   
 Type   
 Description  
smap_ap.coord-1
vector
the coordinates of the points in the low dimensional space  This is the 1th of these quantities
smap_ap.coord-2
vector
the coordinates of the points in the low dimensional space  This is the 2th of these quantities: ARRANGE_POINTSArrange points in a low dimensional space so that the (transformed) distances between points in the low dimensional space match the dissimilarities provided in an input matrix. More details ARGthe initial positions for the projections=smap_mds-1,smap_mds-2 MINTYPE the method to use for the minimisation=smacof TARGET1the matrix of target quantities that you would like to match=smap_hdmat FUNC1a function that is applied on the distances between the points in the low dimensional space={SMAP R_0=2 A=2 B=2} WEIGHTS1the matrix with the weights of the target quantities=lwe_mat MAXITER maximum number of optimization cycles for optimisation algorithms=1000 SMACTOL the tolerance for the smacof algorithm=1E-4 SMACREG this is used to ensure that we don't divide by zero when updating weights for SMACOF algorithm=0.001 TARGET2the matrix of target quantities that you would like to match=smap_mds_mat WEIGHTS2the matrix with the weights of the target quantities=lwe_mat
smapThe VSTACK action with label smap calculates the following quantities:
 Quantity   
 Type   
 Description  
smap
matrix
a matrix that contains the input vectors in its columns: VSTACKCreate a matrix by stacking vectors together More details ARGthe values that you would like to stack together to construct the output matrix=smap_ap.coord-1,smap_ap.coord-2
# --- End of included input --- 
DUMPVECTORPrint a vector to a file More details ARGthe labels of vectors/matrices that should be output in the file=smap.*,lwe FILE the file on which to write the vetors=smap

Output components

This action can calculate the values in the following table when the associated keyword is included in the input for the action. These values can be referenced elsewhere in the input by using this Action's label followed by a dot and the name of the value required from the list below.

Name	Type	Keyword	Description
osample	matrix	PROJECT_ALL	the out-of-sample projections

Full list of keywords

The following table describes the keywords and options that can be used with this action

Keyword	Type	Default	Description
NLOW_DIM	compulsory	none	number of low-dimensional coordinates required
ARG	compulsory	none	the matrix of high dimensional coordinates that you want to project in the low dimensional space
HIGH_DIM_FUNCTION	compulsory	none	the parameters of the switching function in the high dimensional space
LOW_DIM_FUNCTION	compulsory	none	the parameters of the switching function in the low dimensional space
CGTOL	compulsory	1E-6	The tolerance for the conjugate gradient minimization that finds the projection of the landmarks
MAXITER	compulsory	1000	maximum number of optimization cycles for optimisation algorithms
NCYCLES	compulsory	0	The number of cycles of pointwise global optimisation that are required
BUFFER	compulsory	1.1	grid extent for search is (max projection - minimum projection) multiplied by this value
CGRID_SIZE	compulsory	10	number of points to use in each grid direction
FGRID_SIZE	compulsory	0	interpolate the grid onto this number of points -- only works in 2D
OS_CGTOL	compulsory	1E-6	The tolerance for the conjugate gradient minimization that finds the out of sample projections
SMACTOL	compulsory	1E-4	the tolerance for the smacof algorithm
SMACREG	compulsory	0.001	this is used to ensure that we don't divide by zero when updating weights for SMACOF algorithm
WEIGHTS	optional	not used	a vector containing the weights of the points
PROJECT_ALL	optional	false	if the input are landmark coordinates then project the out of sample configurations
USE_SMACOF	optional	false	find the projection in the low dimensional space using the SMACOF algorithm

References

More information about how this action can be used is available in the following articles:

• M. Ceriotti, G. A. Tribello, M. Parrinello, Simplifying the representation of complex free-energy landscapes using sketch-map. Proceedings of the National Academy of Sciences. 108, 13023–13028 (2011)
• G. A. Tribello, M. Ceriotti, M. Parrinello, Using sketch-map coordinates to analyze and bias molecular dynamics simulations. Proceedings of the National Academy of Sciences. 109, 5196–5201 (2012)
• M. Ceriotti, G. A. Tribello, M. Parrinello, Demonstrating the Transferability and the Descriptive Power of Sketch-Map. Journal of Chemical Theory and Computation. 9, 1521–1532 (2013)
• A. Ardevol, G. A. Tribello, M. Ceriotti, M. Parrinello, Probing the Unfolded Configurations of a β-Hairpin Using Sketch-Map. Journal of Chemical Theory and Computation. 11, 1086–1093 (2015)
• A. Ardevol, F. Palazzesi, G. A. Tribello, M. Parrinello, General Protein Data Bank-Based Collective Variables for Protein Folding. Journal of Chemical Theory and Computation. 12, 29–35 (2015)