Today, I released a video overview of DSSR (http://docs.x3dna.org/dssr-overview/).
DSSR has a sizable user base. However, in my opinion, DSSR is still underutilized for what it has to offer. This overview video is intended not only to attract new DSSR users, but also to highlight features that even experienced users may overlook.
As documented in the Overview PDF, DSSR can be easily incorporated into other structural bioinformatics pipelines. Working with Robert Hanson and Thomas Holder respectively, I initiated the integrations of DSSR into Jmol and PyMOL, two of the most popular molecular viewers. The DSSR-Jmol and DSSR-PyMOL integrations lead to unparalleled search capabilities and innovative visualization styles of 3D nucleic acid structures. They also exemplify the critical roles that a domain-specific analysis engine may play in general-purpose molecular visualization tools.
On January 27, 2016, I wrote the blogpost Integrating DSSR into Jmol and PyMOL. Four years later, these integrations have led to two peer-reviewed articles, both published in Nucleic Acids Research (NAR). This blogpost (dated 2020-09-15) highlights key features in each case and reflects on my experience in these two exciting collaborations.
The DSSR-Jmol integration
Hanson RM and Lu XJ (2017). DSSR-enhanced visualization of nucleic acid structures in Jmol. The DSSR-Jmol integration excels in its SQL-like, flexible searching capability of structural features, as demonstrated at the website http://jmol.x3dna.org. This work fills a gap in RNA structural bioinformatics by enabling deep analyses and SQL-like queries of RNA structural characteristics, interactively. Here are some simple examples:
SELECT WITHIN(dssr, "nts WHERE is_modified = true") # modified nucleotides
SELECT pairs # all pairs
Select WITHIN(dssr, "pairs WHERE name = 'Hoogsteen'") # Hoogsteen pairs
SELECT WITHIN(dssr, "pairs WHERE name != 'WC'") # non-Watson-Crick pairs
SELECT junctions # all junctions loops
select within(dssr, "junctions WHERE num_stems = 4") # four-way junction loops
In a recently email communication, Bob wrote:
How are you doing? I’m smiling, because I am remembering our incredible, animated discussions and how fun it was to work together with you on Jmol and DSSR.
The DSSR-PyMOL integration
Lu XJ (2020). DSSR-enabled innovative schematics of 3D nucleic acid structures with PyMOL. The DSSR-PyMOL integration brings unprecedented visual clarity to 3D nucleic acid structures, especially for G-quadruplexes. The four interfaces cover virtually all conceivable use cases. The easiest way to get started and quickly benefit from this work is via the web application at http://skmatic.x3dna.org.
I approached Thomas to write the DSSR-PyMOL manuscript together, in a similar way as the DSSR-Jmol paper. He wrote back, saying “I’m not interesting in being co-author of the paper”, adding:
But, if there is anything I can help you with, like revising the `dssr_block.py` script, or proof-reading the PyMOL related parts of the manuscript, I’ll be happy to do so.
Indeed, Thomas helped in several aspects of the DSSR-PyMOL project, as acknowledged in the paper:
I appreciate Thomas Holder (PyMOL Principal Developer, Schrödinger, Inc.) for writing the DSSR plugin for PyMOL, and for providing insightful comments on the manuscript and the web application interface.
Enhanced vs Innovative
Some viewers may noticed the difference in titles of the two NAR papers: “DSSR-enhanced visualization of nucleic acid structures in Jmol” vs. “DSSR-enabled innovative schematics of 3D nucleic acid structures with PyMOL”. As a matter of fact, the initial title of the DSSR-PyMOL paper was DSSR-enhanced visualization of nucleic acid structures in PyMOL, as shown in the December 02, 2019 announcement post on the 3DNA Forum.
In an era where reproducibility of “scientific” publications has become an issue and “break-throughs” are often broken or hardly held, I hesitate to use phrases such as “innovative”, “novel”, “paradigm shift” etc. Instead, I often use the modest words “refinement”, “enhance”, “improved”, “revised” etc, and try to deliver more than claimed. However, reviewers may take solid work but modest writing as “incremental” or “unexciting”. Before submitting the DSSR-PyMOL paper, I changed the title to DSSR-enabled innovative schematics of 3D nucleic acid structures with PyMOL. Does it mean that the DSSR-PyMOL integration is more innovative than the DSSR-Jmol case? Not necessarily. I do have a paper with “innovative” in its title.
Recently, while reading the Miskiewicz et al. review article How bioinformatics resources work with G4 RNAs, I noticed the term DSSR-G4DB under the category
Databases with G4-related data. It refers to the website http://G4.x3dna.org (or
g4.x3dna.org) that has been there since 2017 and weekly updated with new G-quadruplexes from the PDB. The DSSR-G4 resource, DSSR-Enabled Automatic Identification and Annotation of G-quadruplexes in the PDB, has already been cited several times in literature. However, I have not written up a paper on it yet, and thus have never thought carefully on a name for the resource. The term DSSR-G4DB sounds good to me, and I may well use it in the future.
Given below are the relevant quotations on DSSR and the DSSR-G4DB resource in the Miskiewicz et al. review article and my notes. The underlined headings (e.g., “Conclusion”) are those of the Miskiewicz et al. review article.
Methods: Databases with G4-related data
Currently, there exist 16 databases, which store information concerning quadruplexes. They fall into three categories: databases that collect primary or tertiary structures with experimentally verified G4s (DSSR-G4DB, G4IPDB, G4LDB, G4RNA, Lit392 and Lit638); databases storing data from high-throughput sequencing with mapped quadruplexes (GSE63874, GSE77282, GSE110582 and GSE129281); and databases of sequences with G4s identified in silico (Greglist, GRSDB2, G4-virus, Non-B DB v2.0, Plant-GQ and QuadBase2)
DSSR-G4DB  contains quadruplex nucleic acid structures found by DSSR in the Protein Data Bank , currently 354 entries. The data are annotated. Users can find information about G-tetrads, G4 helices and G4-stems and visualize the 3D models of G4 structures. Availability: webserver (http://g4.x3 dna.org). Recent update: 5 June 2020.
Note: DSSR-G4DB is updated weekly. The latest update is on 2020-09-09, with 362 G-quadruplexes auto-curated with DSSR from the PDB.
Methods: Tools that analyze and visualize 2D and 3D structure
Currently, four tools can analyze and visualize G4 structures. DSSR  … ElTetrado  … RNApdbee [66, 69] … 3D-NuS 
DSSR  processes the 3D structure of the RNA molecule and annotates its secondary structure. It is a part of the 3DNA suite  designed to work with the structures of nucleic acids. DSSR identifies, classifies and describes base pairs, multiplets and characteristic motifs of the secondary structure; helices, stems, hairpin loops, bulges, internal loops, junctions and others. It can also detect modules and tertiary structure patterns, includ- ing pseudoknots and kink-turns. The recent extension, DSSR- PyMOL , allows drawing cartoon-block schemes of the 3D structure and responds to the need for simplified visualization of quadruplexes. Input data formats: PDB, mmCIF and PDB ID. Availability: standalone program, web application (http://dssr.x3 dna.org/, http://skmatic.x3dna.org/).
Note: The other three tools all depend on or make use of DSSR and 3DNA:
- ElTetrado “ElTetrado depends on DSSR (Lu, Bussemaker and Olson, 2015) in terms of detection of base pairing and stacking.”
- RNApdbee uses 3DNA/DSSR as the default to identify base pairs.
- 3D-NuS employs 3DNA for structural analysis and model building.
“These filtrated structures (225 DNA and 166 RNA structures) have been used to derive the local base pair step and base pair parameters (Table S2 for DNA and Table S3 for RNA) using 3DNA software package  and are stored in the server for 3D-NuS modeling.”
“Soon after the user submits input for sequence-specific modeling, the server fetches the appropriate base pair step and base pair parameters from the database and creates a 3DNA style input file. Subsequently, the template model is built using the rebuild module of 3DNA software package and subjected to energy optimization using X-plor  to remove steric hindrance, specifically in the mismatch- containing duplexes (Fig. 1).”
Results: Computational experiments with structure-based tools
DSSR and ElTetrado identified quadruplexes in the input PDB files. Both programs focused on structural aspects of the input molecule, explicitly informing about quadruplexes and tetrads within the structure. DSSR provided an extensive analysis of 3D structures and output the data about G-tetrads, G-helices and G4-stems. It computed planarity for each G-tetrad and gave the sections area, rise and twist parameters for G4-helix and G4-stems. The program automatically assigned loop topologies according to the predefined types (P—parallel, D—diagonal and L—lateral) and their orientation (+/−). DSSR-PyMOL generated block schemes of both quadruplexes (Figure 4A3 and B3). ElTetrado also calculated planarity, rise and twist parameters and identified strand directions for both quadruplexes. It classified the quadruplexes and their component tetrads to ONZ classes. Finally, it generated the arc diagram (Figure 4A1 and B1) and two-line dot-bracket encoding of every quadruplex.
Note: DSSR contains an undocumented option
--G4. With the
ONZ variant, i.e.,
--g4=onz (case does not matter), DSSR also outputs the ONZ classification of G-tetrads from the same chain.
DSSR comprehensively examines the G4 structure, determines a variety of its parameters and provides the schematic 3D view.
It is worth noting that DSSR has been categorized under “Databases with G4-related data” and “Tools that analyze and visualize 2D and 3D structure” of the Methods section. It is not a tool that predicts G4 location in the sequence. There are 14 tools listed in “Table 2. Selected features of PQS prediction tools”, including G4Hunter and QGRS Mapper etc.
Recently, while visiting the NAR website on DSSR-enabled innovative schematics of 3D nucleic acid structures with PyMOL, I noticed a big red circle ① near “View Metrics”. I was quite curious to see what it meant. After a few clicks, I was delighted to read the following recommendation in Faculty Opinions by Quentin Vicens:
I really enjoyed “playing” with the revised and expanded version of Dissecting the Spatial Structure of RNA (DSSR) described by Xiang-Jun Lu in this July issue of NAR. The software is known to generate ‘block view’ representations of nucleic acids that make many parameters more immediately visible, such as base composition, stacking, and groove depth. This new version includes Watson-Crick pairs shown as single rectangles, and G quadruplexes as large squares, making such regions more quickly distinguishable from other regions within an overall tertiary structure. I was amazed at how simple and effective the web interface was, and I liked the possibility to download a PyMOL session to look at molecules under different angles. If need be, blocks can be further edited in PyMOL using the provided plugin (see on page 35). I highly recommend it!
The DSSR-PyMOL schematics paper/website has been rated “Very Good”, and classified as “Good for Teaching”. See Vicens Q: Faculty Opinions Recommendation of [Lu XJ, Nucleic Acids Res 2020 48(13):e74]. In Faculty Opinions, 14 Aug 2020; 10.3410/f.738001682.793577327.
DSSR 2.0 is out. It integrates an unprecedented set of features into one computational tool, including analysis/annotation, schematic visualization, and model building of 3D nucleic acid structures. DSSR 2.0 supersedes 3DNA 2.4, which is still maintained but no additional features other than bug fixes are scheduled. See the DSSR 2.0 overview PDF.
DSSR delivers a great user experience by solving problems and saving time. Considering its usability, interoperability, features, and support, DSSR easily stands out among `competitors’. It exemplifies a `solid software product’. I strive to make DSSR a pragmatic tool that the structural bioinformatics community can count on.
DSSR 2.0 is licensed by Columbia University. The software remains free for academic users, with the basic user manual. The professional user manual (over 230 pages, including 7 appendices) is available for paid academic users or commercial users only. Licensing revenue helps ensure the long-term sustainability of the DSSR project.
Additionally, the paper “DSSR-enabled innovative schematics of 3D nucleic acid structures with PyMOL” has recently been published in Nucleic Acids Research, 48(13):e74. Check the web interface.
The DSSR-PyMOL paper/website has been rated “very good” and classified as “Good for Teaching”. See Vicens Q: Faculty Opinions Recommendation of [Lu XJ, Nucleic Acids Res 2020 48(13):e74]. In Faculty Opinions, 14 Aug 2020; 10.3410/f.738001682.793577327
Recently I performed a survey of citations to thirteen 3DNA-related publications using Web of Science from Clarivate Analytics. The time range is from 2015 to 2020 (June 30), for a total of five-and-half years. The 1,050 citations span 223 scientific journals, covering a broad range of research fields such as biology, medicine, chemistry, physics, materials etc. Not surprisingly, the citing journals include Cell, Nature and sub-journals, Science, and PNAS.
Each of following six papers has been cited over 50 times, as detailed below. Adding the six numbers together, there are 962 citations, accounting for 92% of the total 1,050.
402 times in 138 journals] Lu,X.-J. and Olson,W.K. (2003) 3DNA: A software package for the analysis, rebuilding and visualization of three-dimensional nucleic acid structures. Nucleic Acids Res., 31, 5108–5121.
201 times in 81 journals] Lu,X.-J. and Olson,W.K. (2008) 3DNA: A versatile, integrated software system for the analysis, rebuilding and visualization of three-dimensional nucleic-acid structures. Nat. Protoc., 3, 1213–1227.
127 times in 71 journals] Zheng,G., Lu,X.J. and Olson,W.K. (2009) Web 3DNA––a web server for the analysis, reconstruction, and visualization of three-dimensional nucleic-acid structures. Nucleic Acids Res, 37, W240-6.
115 times in 57 journals] Olson,W.K., Bansal,M., Burley,S.K., Dickerson,R.E., Gerstein,M., Harvey,S.C., Heinemann,U., Lu,X.-J., Neidle,S., Shakked,Z., Sklenar,H., Suzuki,M., Tung,C.-S., Westhof,E., Wolberger,C. and Berman,H.M. (2001) A standard reference frame for the description of nucleic acid base-pair geometry. J. Mol. Biol., 313, 229–237.
66 times in 32 journals] Lu,X.-J., Bussemaker,H.J. and Olson,W.K. (2015) DSSR: An integrated software tool for dissecting the spatial structure of RNA. Nucleic Acids Res., 43, e142.
51 times in 41 journals] Lu,X.J., Shakked,Z. and Olson,W.K. (2000) A-form conformational motifs in ligand-bound DNA structures. J. Mol. Biol., 300, 819–40.
The top 21 journals that cite 3DNA papers 10 times or more are listed below. Nucleic Acids Research stands out, with a total of 148 citations, accounting for 14% of the total 1,050 citations.
148 Nucleic Acids Research
84 Journal of Physical Chemistry B
40 Physical Chemistry Chemical Physics
34 Biophysical Journal
33 Journal of Chemical Theory and Computation
24 PLoS One
24 Scientific Reports
22 Journal of Biomolecular Structure & Dynamics
19 Journal of Chemical Information and Modeling
16 Nature Communications
15 Journal of the American Chemical Society
12 Acta Crystallographica Section D: Structural Biology
12 Journal of Molecular Modeling
11 Chemistry: a European Journal
11 Journal of Chemical Physics
10 Journal of Biological Chemistry
Following my previous post 3DNA/blocview-PyMOL images in covers of the RNA journal in 2019, here is an update for 2020. The cover images of the January to July issues have all been generated with help of 3DNA and provided by the NDB:
RNA is displayed as a red ribbon; block bases use NDB colors: A—red, C—yellow, G—green, U—cyan. The image was generated using 3DNA/blocview and PyMol software. Cover image provided by the Nucleic Acid Database (ndbserver.rutgers.edu).
Here is the composite figure of the seven cover images, with the brand new DSSR-PyMOL schematics for comparison.
Details of the seven structures illustrated in the cover images are described below:
- January 2020 Pumilio homolog PUF domain in complex with RNA (PDB id: 5yki; Zhao YY, Mao MW, Zhang WJ, Wang J, Li HT, Yang Y, Wang Z, Wu JW. 2018. Expanding RNA binding specificity and affinity of engineered PUF domains. Nucleic Acids Res 46: 4771–4782). Engineered nine-repeat PUF domain binds to its RNA target specifically and with high binding affinity.
- February 2020 Aprataxin RNA–DNA deadenylase product complex (PDB id: 6cvo; Tumbale P, Schellenberg MJ, Mueller GA, Fairweather E, Watson M, Little JN, Krahn J, Waddell I, London RE, Williams RS. 2018. Mechanism of APTX nicked DNA sensing and pleiotropic inactivation in neurodegenerative disease. EMBO J 37: e98875). Human aprataxin RNA–DNA deadenylase protects genome integrity and corrects abortive DNA ligation arising during ribonucleotide excision repair and base excision DNA repair.
- March 2020 PreQ1 riboswitch (PDB id: 6e1w; Connelly CM, Numata T, Boer RE, Moon MH, Sinniah RS, Barchi JJ, Ferre-D’Amare AR, Schneekloth Jr JS. 2019. Synthetic ligands for PreQ1 riboswitches provide structural and mechanistic insights into targeting RNA tertiary structure. Nat Commun 10: 1501). Class I PreQ1 riboswitch regulates downstream gene expression in response to its cognate ligand PreQ1 (7-aminomethyl-7-deazaguanine).
- April 2020 Hatchet ribozyme (PDB id: 6jq6; Zheng L, Falschlunger C, Huang K, Mairhofer E, Yuan S, Wang J, Patel DJ, Micura R, Ren A. 2019. Hatchet ribozyme structure and implications for cleavage mechanism. Proc Natl Acad Sci 116: 10783–10791). This crystal structure of the hatchet ribozyme product features a compact symmetric dimer.
- May 2020 Adenovirus virus-associated RNA (PDB id: 6ol3; Hood IV, Gordon JM, Bou-Nader C, Henderson FE, Bahmanjah S, Zhang J. 2019. Crystal structure of an adenovirus virus-associated RNA. Nat Commun 10: 2871). Acutely bent viral RNA fragment is a protein kinase R inhibitor and features an unusually structured apical loop, a wobble-enriched, coaxially stacked apical and tetra-stems, and a central domain pseudoknot that resembles codon-anticodon interactions.
- June 2020 Archeoglobus fulgidus L7Ae bound to cognate K-turn (PDB id: 6hct; Huang L, Ashraf S, Lilley DMJ. 2019. The role of RNA structure in translational regulation by L7Ae protein in archaea. RNA 25: 60–69). 50S archaeal ribosome protein L7Ae binds to a K-turn structure in the 5′-leader of the mRNA of its structural gene to regulate translation.
- July 2020 Spinach RNA aptamer/Fab complex (PDB id: 6b14; Koirala D, Shelke SA, Dupont M, Ruiz S, DasGupta S, Bailey LJ, Benner SA, Piccirilli JA. 2018. Affinity maturation of a portable Fab-RNA module for chaperone-assisted RNA crystallography. Nucleic Acids Res 46: 2624–2635). Novel Fab-RNA module can serve as an affinity tag for RNA purification and imaging and as a chaperone for RNA crystallography.
The paper, titled DSSR-enabled innovative schematics of 3D nucleic acid structures with PyMOL, has just been published in Nucleic Acids Research (online on May 22, 2020). Here is the abstract:
Sophisticated analysis and simplified visualization are crucial for understanding complicated structures of biomacromolecules. DSSR (Dissecting the Spatial Structure of RNA) is an integrated computational tool that has streamlined the analysis and annotation of 3D nucleic acid structures. The program creates schematic block representations in diverse styles that can be seamlessly integrated into PyMOL and complement its other popular visualization options. In addition to portraying individual base blocks, DSSR can draw Watson-Crick pairs as long blocks and highlight the minor-groove edges. Notably, DSSR can dramatically simplify the depiction of G-quadruplexes by automatically detecting G-tetrads and treating them as large square blocks. The DSSR-enabled innovative schematics with PyMOL are aesthetically pleasing and highly informative: the base identity, pairing geometry, stacking interactions, double-helical stems, and G-quadruplexes are immediately obvious. These features can be accessed via four interfaces: the command-line interface, the DSSR plugin for PyMOL, the web application, and the web application programming interface. The supplemental PDF serves as a practical guide, with complete and reproducible examples. Thus, even beginners or occasional users can get started quickly, especially via the web application at http://skmatic.x3dna.org.
A brief history on DNA/RNA schematics as implemented in SCHNAaP/SCHNArP, 3DNA, and now in DSSR:
The idea of representing bases and WC-pairs as rectangular blocks came from the pioneering work of Calladine et al. (27,28) The block schematics were first implemented in the pair of SCHNAaP/SCHNArP programs (29,30) for rigorous analysis and reversible rebuilding of double-helical nucleic acid structures. The algorithms that underpinned SCHNAaP/SCHNArP laid the foundation of ‘analyze’ and ‘rebuild’, two core components of the 3DNA suite of programs (31–33). 3DNA also takes advantage of the standard base reference frame (34), and comprises quite a few other related programs. One of them is ‘blocview’, a script which calls several 3DNA utility programs to generate individual base blocks and set the view, MolScript (35) to produce backbone ribbons, and Raster3D (36) to render the composite image. The 3DNA ‘blocview’ schematics catch characteristic attributes of nucleic acid structures. They have gradually become popular and been adopted into the RCSB PDB (1) and the NDB (37), and then propagated into other bioinformatics resources (e.g., the ‘RNA Structure Atlas’ website hosted by the Leontis-Zirbel RNA group).
DSSR supersedes ‘blocview’ by eliminating all the internal and external dependencies of the 3DNA utility program. DSSR produces block representations, not only of individual bases but also WC-pairs and G-tetrads, that can be fed directly into PyMOL. The DSSR-PyMOL integration is easier to use, has more features, and produces better schematics than the original 3DNA-blocview approach.
Indeed, the base block schematics have continuously evolved for over two decades, as appreciated in the acknowledgements:
I would like to thank Christopher A. Hunter, Christopher R. Calladine, Helen M. Berman, Catherine L. Lawson, Zukang Feng, Wilma K. Olson and Harmen J. Bussemaker for their helpful input on the block schematic during its continuous evolution for over two decades. I appreciate Thomas Holder (PyMOL Principal Developer, Schrödinger, Inc.) for writing the DSSR plugin for PyMOL, and for providing insightful comments on the manuscript and the web application interface. I also thank Jessalyn Lu and Yin Yin Lu for proofreading the manuscript, and the user community for feedback.
Notably, the supplemental PDF has been diligently written to serve as a practical guide, with complete and reproducible examples. In fact, the paper concludes with the following two sentences:
Finally, all results reported here are completely reproduceable (see the supplemental PDF). Any questions related to this work are welcome and will be openly addressed on the 3DNA Forum (http://forum.x3dna.org).
As of version 2.0 (to be released soon), DSSR has a new module for in silico base mutations that is context sensitive. Powered by the DSSR analysis engine, the module allows users to perform base mutations in unprecedented flexibility and convenance. Here are some examples:
- Mutate all bases in hairpin loops to a specific base (e.g., G)
- Mutate all non-stem bases to a specific base (e.g., U)
- Mutate bases 2-12 to a specific base (e.g., A) regardless of context
- Mutate bases 1-10 in a given structure to a new sequence (e.g., AUAUAUAUAU)
- Mutate all bases of the same type to another (e.g., A to G)
- Mutate all bases of the same type to another (e.g., C to U) except for some nucleotides
- Mutate all G-C Watson-Crick (WC) pairs to C-G WC pairs, and A-U to U-A
- Mutate all G-tetrads in G-quadruplexes to non-G-tetrads (e.g., U-tetrads)
By default, the mutation preserves both the geometry of the sugar-phosphate backbone and the base reference frame (position and orientation). As a result, re-analyzing the mutated model gives the same base-pair and step parameters as those of the original structure.
Over the years, the 3DNA
mutate bases program has been cited in the literature and patent, including the following ones:
- Howe, John A., et al. Selective small-molecule inhibition of an RNA structural element. Nature 526.7575 (2015): 672-677.
- Wang, Hao, et al. Dual-targeting small-molecule inhibitors of the Staphylococcus aureus FMN riboswitch disrupt riboflavin homeostasis in an infectious setting. Cell Chemical Biology 24.5 (2017): 576-588.
- AlQuraishi, Mohammed, and Harley H. McAdams. Three enhancements to the inference of statistical protein‐DNA potentials. Proteins: Structure, Function, and Bioinformatics 81.3 (2013): 426-442.
- Wang, Harris, Sagi Shapira, and Victoria Stockman. High-throughput strategy for dissecting mammalian genetic interactions. U.S. Patent Application No. 15/747,677.
The DSSR mutation module has completely obsoleted the
mutate_bases program distributed in 3DNA v2.x. In addition to serving as a drop-in replacement of
mutate_bases, the DSSR approach offers much more features and versatility: it is simply better.