HERMES is written in MATLAB script language. The MATLAB functions that comprise HERMES, their interconnections and their purpose are described in the flow chart below
the structural model of the nucleic acid has to be submitted in PDB format. For more information, visit the corresponding PDB documentation. In case of the Assess multiple models application, the user can submit more than two models. If only one structural model is to be submitted, the user is encouraged to use the Assess experimental fiRDCs application.
“Chains”
it refers to the chains the user is interested in. They have to be submitted exactly in the way they appear in the structural model. In case of submitting more than one chain, the letters of the chains should be written without any (space) character in between, e.g. AB
“List of RDCs to be predicted”, “Experimental RDCs” (fiRDC_input)
Depending of the the application the user selects, HERMES will require either the list of the interacting nuclei for which field-induced RDCs are to be calculated (Predict fiRDCs application) or the list of the experimentally acquired field-induced RDCs (for the rest of the applications). The last columns of the
file should contain the generalized order parameter S for each pair of interacting nuclei. This dimensionless parameter provides a measure of the spatial amplitude of the motion of the vector of two nuclei that participate in a chemical bond. Higher values of S describe more rigid chemical bonds (S = 1 is interpreted to no internal motion, while S = 0 refers to completely isotropic internal motion of the chemical bond). The value of S is defined for each pair of interacting nuclei in the last column of the user-submitted fiRDCs template. There is an option of commenting line/excluding experimental values by adding the "#" symbol in the beginning of the line, as shown below :
Column
Content
1
chain
4-6
nucleotide serial number
9-12
interacting atom 1
14-17
interacting atom 2
20-26
value of experimental fiRDCs
30-35
experimental error
38-42
generalized order parameter S
“Bhigh (in T)”, “Blow (in T)”
the magnetic induction values should be given in Tesla. B high and B low refer to “high field” and “low field” values for which the NMR experiment was conducted. Below a handy MHz-to-T conversion tool is given (100 MHz -> 2.37 T):
Spectrometer frequency: MHz
Magnetic Field Induction: Tesla
“Temperature (in K)”
The temperature of the experiment should be given in Kelvin.
“List of nucleotides for domain 1, domain 2”
the two sets of nucleotides that belong, respectively, to each one of the two domains. The set that is defined as Domain 1 remains in the same place during the calculations, and all rotations are performed on the nucleotides that comprise Domain 2.
“Angle increment for theta, phi”
the increment (in degrees) between each value for the polar angles theta, phi that describe the relative orientation of the user-defined domains in the Relative Domain Orientation Application
it refers to the coordinate system of each base type (base reference), given in PDB format. It should include the coordinates of the nuclei that belong to the ring(s) of each base type that appears in the structural model. The expected format and the default geometries for A, G, C, T, U bases are given below :
Column
Content
1-6
'ATOM'
13-16
atom name
18-20
nucleotide name
22
chain
31-38
x coordinate
39-46
y coordinate
47-54
z coordinate
“Magnetic susceptibility tensor”
The user has the option to submit their own base reference and they are prompted to give their own magnetic susceptibility tensor for each base type.
The user can also submit their own 3-letter code for a modified base and the corresponding magnetic susceptibility tensor. The 3-letter code should also correspond to the structural PDB model and
to the experimental fiRDCs file. The default value for bases is a 3x3 symmetric matrix resulting from DFT/GIAO calculations in Gaussian 09 :
Each application produces a compressed (.zip) file that the user can download, which contains the outputs according to the application (refer to Table). This file contains all the ouputs, namely :
“Euler angles (Euler_angles_out)”
A set of eight angles (due to the four-fold degeneracy of the alignment tensor) that relates the (cartesian) reference system of the molecule to the principal axis system of the molecular alignment tensor is stored in a tab-delimited file. The calculation of Euler angles follows the y-convention, that is a sequence of three consecutive elemental rotations about the mobile (z, y', z'') axis of the molecular coordinate system.
“Tait-Bryan angles (Tait_Bryan_angles_out)”
A set of eight angles (due to the four-fold degeneracy of the alignment tensor) that relates the reference system of the molecule (cartesian) to the principal axis system of the molecular alignment tensor is stored in a tab-delimited file. The calculation of Tait-Bryan angles is based on a sequence of three consecutive elemental rotations about three different fixed axes mobile (x, y, z). The output, either for Euler angles or for Tait-Bryan angles, rearranged in a spreadsheet, has the following format:
V11, V22, V33 are the three column vectors that comprise the (3x3) molecular alignment tensor. Due to its four-fold degeneracy, we report four tensors (V1, V2, V3, V4), each one having its own set of three angles (a, b, c) that describe the rotations. More accurately, we have :
Tensor
Angles, set 1
Angles, set 2
V1
a, b, c
a+180, -b, c+180
V2
a, b, c+180
a+180, -b, c
V3
a, b+180, -c+180
a+180, 180-b, -c
V4
a, b+180, -c
a+180, -b, -c+180
Set 1 and set 2 of angles are equivalent for each alignment tensor. Therefore, every six entries of the produced file correspond to two sets of angles for one alignment tensor. As for the above example, set 1 for V1 is (a, b, c)=(74.2094695259968, 97.7695481010489, 88.4079720574698)
“Molecular alignment tensor (PDB_out)”
The submitted structural model is enriched with its molecular alignment tensor, calculated by HERMES. The coordinates of the tensor are placed in the last lines of the produced .pdb file. In case of the Relative Domain Orientation application, two alignment tensors are calculated, one for each domain.
In the example above, the last lines of the output structural model contain the coordinates (x, y, z) of the alignment tensor of a molecule. The tensor can be visualized with the molecular graphics sofware PyMOL.
“Predicted fiRDCs (fiRDCs_out)”
The values of the predicted fiRDCs for each pair of interacting nuclei are stored in the last column of the user-submitted template.
“Statistical Analysis”
HERMES performs statistical analysis of the data according to the selected application, which is displayed on the interface and is also available for downloading.
Scatter plot of the predicted fiRDCs (plot_out)
Simple linear regression plot and Deming regression plot between experimental and predicted fiRDCs (corr_plot_out)
RMSD (root-mean-square deviation)
Pearson's correlation coefficient
Q- and R-factors
ANOVA test p-value, Tukey HSD post-hoc p-values
The above mentioned statistical descriptors are listed in the output files "Statistics.txt" and "TukeyHSD.txt"
“Structural models with different relative orientations”
A set of structural models with different relative domain orientations is produced according to the user-defined increments for polar angles. This applies only for the Relative Domain Orientation application. The output models are in standard PDB format.