DRAWNA: A program for drawing schematic views of nucleic acids C. Massire,* C. Gaspin*? and E. Westhof* *Equipe de ~od~lisation et de Simulation des Acides Nu~l~~ques, UPR Structure des macromolecules ~~olog~ques et Mecanismes de Reconnaissance, Institut de Biologic Moleculaire et Cellulaire du CNRS, France i-Station de Biome’trie et d’lntelligence Artifcielle, Institut National de la Recherche Agronomique, France A program for drawing automatically exact and schematic views of nucleic acids is described. Th.e program is written in C ANSI and uses the Silicon Graphics GL and Xirisw libraries within the X1 IMottfenvironment. Through menus, the user can choose, specify, and manipuLate in real time the three-dimensional views to be displayed. drawing options include partitioning of structures into differently colored or shaped fragments, representation of backbones as flat or with conic-section ribbons, display of paired or free bases as rods, and display of surjaces as zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA filled or outlined and stereo or depth-cued views. Keywords: nucleic acids, three-dimensional representation, GL, Motif; B-spline INTRODUCTION necessary to visualize and study three-dimensional (3D) folding, i.e., the relative spatial organization of molecular components. To do so, molecular biologists need to have at their disposal computer graphics programs able to give rapid and exact schematic views of 3D structures within a reason- able time and without extensive computer know-how. Such programs, displaying 30 representations of biological mac- romolecules are now available, e.g., RIBBONS,3s” MOLSCRIPT,5 and SETOR, among others. These pro- grams are convenient for drawing proteins and sometimes may also be applied to nucleic acids. This is usually done by renaming phosphorus atoms as C, in the input coordinate file. However, their schematic representations of structural subunits (cylinders, P-sheets, etc.) are inadequate for draw- ing base pairs in the case of nucleic acids. The program RIBBONS4 produces beautiful drawings of nucleic acids, but requires a cumbersome number of separate files prevent- ing an automation of the drawing process. Nucleic acids are characterized by the sequence of bases attached on the polynucleotide sugar-phosphate backbone (normally numbered in the 5’ to 3’ direction). In DNA, the main structure is the antiparallel double-stranded helix maintained via hydrogen bonds involving atoms of Watson- Crick complementary bases (A. . . T and G . . . C). Except in some viruses where it serves as genetic material, RNA is generaliy single-stranded, and folding of the polynucleotide sugar-phosphate backbone allows the bases to interact and form hydrogen-bonded double-stranded helices separated by single-stranded regions. The formed elements of secondary structure (helices, loops, or bulges) interact with each other to stabilize the tertiary structure, which can be as intricate and complex as in proteins.‘,2 For understanding the function as well as evolutionary and mutational data of biological macromolecules, it is We wish to present a new program, DRAWNA, specifi- cally devoted to fast and automatic drawing of 3D structures of ribonucleic (RNA) or deoxyribonucleic (DNA) acid structures with the possibility of displaying at the same time biological or chemical data. Besides atomic and detailed molecular representations, schematic drawings constitute important mnemotechnic and heuristic tools. Indeed, like the sketches of humorists, schematic drawings outline the rele- vant and salient features compared with the contingent ap- pearances. However, the usefulness of a schematic drawing is directly correlated with its accuracy; i.e., the detailed and schematic representations should be congruent. This twofold requirement (“objective drawings of subjective rep- resentations”) leads to potential ambiguities since the use- fulness proceeds from the partiality in the choice of the emerging characteristics.7 The choices made should there- fore be explicitly stated, along with their limits and bias. In this way, iconography can reach a status other than merely that of being supplementary material. Color Plates for this article are on page 196. Received 2 t December 1993; revised 18 January 1994; accepted 25 January 1994 Address reprint requests to Professor Westhof at Equipe de Modtlisation et de Simulation des Acides Nucldiques, UPR Structure des Macromolecules Biologiques et Mecanismes de Reconnaissance, Institut de Biologie Moltculaire et Cellulaire du CNRS, 15, rue Rend Descartes, 67084 Stras- bourg-CCdex, France. Nowadays, RNA folding is viewed as resulting from the compaction of preformed two-dimensional (2D) compo- nents in 3D space.’ According to this hierarchical view, helices and hairpin loops form first. Helices then interact locally end-to-end through stacking or by forming pseudoknots or triple helices. Finally, these autonomously folded subdomains associate cooperatively by loop-loop 0 1994 Butterworth-Heinemann .I. Mol. Graphics, 1994, Vol. 12, September 201