An expert that computes everything of RNA

This web-server executes two software for the user for any DNA or RNA three-dimensional structure for analysis of different features of the given structure. Most three-dimensional structures of nucleic acids are stabilized by many hydrogen bonded base pairs. These base pairs can be of canonical Watson-Crick type of non-canonical type. Most of the nucleic acids are solved by X-ray crystallography, where hydrogen atom positions are unavailable but hydrogen atomic positions are crucial for detection of hydrogen bonds stabilizing the base pairs and the overall structure. The BPFIND identifies the base pairs using a hypothesis driven approach without explicit consideration of the hydrogen atomic positions. The outputs of BPFIND are internally used by NUPARM for analysis of all the structural parameters – starting from torsion angles to intra- and inter-base pair parameters, base-base proximity, stacking overlap, etc.


One needs to either (i) provide PDB-ID of an RNA or DNA structure “at the box indicating Enter CIF accn number...” and analyze it or (ii) upload a model from local computer “by using Browse...” button for analysis of its structure. The uploaded structure can be in PDB or mmCif format. Several user-defined options are available for both BPFIND and NUPARM, which is useful for different specific purposes. The default options have been set by the developers, which is suitable for most regular practices. However, alteration of these parameters is often foreseen. The user can click “Change the Default for BPFIND” to verify the Default set options and change any of them, if required.

Hydrogen bond distance cutoff: [The value is set to 3.8 so as to identify most known WatsonCrick and non-canonical base pairs. However, if the user wants to identify only very good hydrogen bonds (s/)he can modify it to smaller value. Increasing it to larger value would identify many more base pairs, some of which may not really be called a base pair.]

Pseudo angle cutoff: [The value is set to 120o, considering consistency with others. Making this value smaller may reduce number of identified base pairs in a structure.]

E-value cutoff: [Larger value, if used, may identify some base pairs of more poor geometry.]

Select desired chain identifier in PDB entry: [Several PDB-IDs have structures of multiple RNA/DNA chains and parameters for all of these chains are often not required. One can enter a specific Chain ID for those files. The default is set so that structures for all the chains are analyzed.]

Select desired NMR model in the File: [Most often a structure solved by NMR spectroscopy provide multiple possible models of the same. By default the server chooses to calculate parameters for the first model. However, if a user is interested to analyze some other model, (s/)he can mention that model number (usually an integer). If one wants to analyze all the models, (s/)he has to run the program several times depending on number of models given in the PDB file.]

Include HETATM entries in PDB: [Often a functional RNA possesses several modified nucleosides, such as PSU in tRNA. By default the server does not consider these residues but one can click that option to include these residues also for calculation of parameters. The database of modified residues can be updated by receiving your feedback]

Avoid identification of base pairs stabilized by C-H...O/N H-bonds: [Hydrogen bonds are generally between a polar D—H group, where D (donor) is usually N or O and a polar Acceptor (A) atom. However, recent results indicate C—H...O/N hydrogen bonds also have stabilizing factor. Hence the The user can click “Change the Default for BPFIND” to verify the Default set options and change any of them, if required program, by default, considers all known types of hydrogen bonds. However, if one is interested in analysis of strongly coupled base pairs, (s/)he can click the option to consider only polar hydrogen bonds.]

Avoid identification of base pairs involving sugar O2' atoms: [Base pairs, in true sense, are by hydrogen bonds between the two bases involved, but in RNA 2'-OH group has big role to play. Hence, by default, BPFIND and NUPARM considers base pairs stabilized by at most one hydrogen bond involving 2'-OH group of the ribose sugar. If the user is not interested to analyze these, (s/)he can click this button to opt out.]

Avoid base pairing between residue no. i and i+1: [In most cases two residues, far from each other in sequence, form base pair in stabilizing the RNA/DNA structures while two consecutive residues, forming base pairs cannot form double helix. Such base pairing between consecutive residues are also termed as Dinucleotide Platform. One can click the option not to consider such Dinucleotide Platforms to be identified.]

Avoid printing base pairing information w.r.t. the second strand.
This is suitable for simple oligonucleotides:
[In case of pure double helical structure, such as many synthetic DNA or RNA structures in PDB, the residues of second strand are obviously paired to appropriate residues of the first strand. Hence, these informations are duplicated. One can click the approach box to suppress such information. One should be careful in this selection as if the structure has unpaired residue, such as in most functional RNA, the selection inappropriately attempts to detect the second strand and calculates parameters of a fragment of the molecule.]

Similarly the user can click “Change the Default for NUPARMto verify the Default set options and change any of them, if required. There are fewer of them.

Do not use default cross-product method: [Finding the vectors normal to each base is vital for calculation of most parameters and none of the methods are perfect. By default the NUPARM program calculates all atom-triads (three atoms at a time) to find out normals to them and calculates average of these normals. Clicking this box forces the program to use Least-Squares method for finding mean base perpendicular.]

Use C1' - C1' as Y-axis: [Traditionally Dickerson's FREEDNA program used to consider the vector joining two C1' atoms of the two paired bases as base pair Y-axis. This has some other limitation as the mid-point between two C1' atoms is far from the center of mass of the base pair. Choice of C8—C6 joining line as base pair Y-axis is better in that respect and that is the default. The user can click this button to compare results with FREEDNA parameters.]

Do not use C6-C8 midpoint as BP center: [One can select this button to use center of mass of a base pair instead of C6—C8 midpoint]

Required Single stranded parameters: [This is often required for complete analysis of DNA/RNA complete double helical molecules but for functional RNA such calculation may not be required.]

Consider all base pairs as Watson-crick type: [NUPARM, by default, calculates all intra-base pair parameters, (Propeller, Buckle, etc.,) considering hydrogen bonding edge specific axis system. As a result parameters of a good non-canonical base pair also are calculated near zero (except for Stretch, which is true distance between the two bases and is around 2.8). This option has been kept to compare NUPARM results, especially for intra-base pair parameters, with parameters calculated by CURVES or 3DNA packages.]