[ Graphics | Simulations | Sequence | Modules ]

[ Views | Skin | RNA/DNA | Combos ]

ICM and ICM-derived plugins provide a viewing environment for large a small molecules and general three-dimensional objects with or without textures. Various types of enhancements including stereo, anti-aliasing, graphical layers, on-the-fly generation of shadows, occlusion shading, custom backgrounds, depth cueing and simple rotation, translation, zooming, clipping, picking, continuous movements, separate

The views include binding and active site surfaces with mapped properties automatic identification and views of cavities and open binding pockets electrostatic surfaces

The contour-buildup algorithm calculates the smooth and accurate analytical molecular surface in seconds. This surface can be saved as a geometrical object, saved as a vectorized postscript file.

* a Gaussian molecular density which can be contoured at different levels and to generate different smooth molecular envelopes and enclosed pockets and cavities: build string "HEK" ; display a_ xstick # tripeptide make map potential Box( a_ 3.) make grob m_atoms exact 0.5 solid display g_atoms smooth transparent

PDB entry: 101d ICM command: nice "103d"

PDB entry: 4tna ICM commands: nice "4tna" color ribbon a_N/* Count(Nof(a_N/*))

Simplified molecular representations are built automatically (e.g. the protein-dna complex is shown with one command: nice "1dnk" ). You can combine different types of molecular representations with solid or wire geometrical objects.

[ Peptides | Homology | Design | Symmetry | Protein-Protein Docking | Pockets | Docking & VLS | 3D chemical editing | Electrostatics ]

Take a peptide sequence and predict its three-dimensional structure. Of course, success is not guaranteed, especially if the peptide is longer than about 25 residues but some preliminary tests are encouraging.

ICM has an excellent record in building accurate models by homology. The procedure will build the framework, shake up the side-chains and loops by global energy optimization. You can also color the model by local reliability to identify the potentially wrong parts of the model.

ICM was used to design two new 7 residue loops and in both cases the designs were successful. Moreover, the predicted conformations turned out to be exactly right (accuracy of 0.5A) after the crystallographic structures of the designed proteins were determined in Rik Wierenga's lab. Use the _loop script to predict loop conformations and calcEnergyStrain to identify the strained parts of the design.

ICM has a database of loops compiled from PDB structures. To search a fragment against the database and load the conformers into a session (eg into a stack), use the following command: build loop stack rs_fragment . The display stack command will allow to view the conformers. You may also load the stack into the object with the load stack a_ command.

ICM has a full set of commands and functions to generate symmetry related molecules and generate "biological units".

As demonstrated in several recent papers, short flexible peptides can be successfully docked ab initio to their receptors. This method is a blend of the peptide folding with the grid potentials representing the receptor. A similar method can be applied to any chemical. A chemical can be built from a 2D representation and optimized. The "druggable" pockets can be predicted with an algorithm based on the contiguous grid energy densities.

Accurate and fast potentials and empirically adjusted scoring functions have led to an efficient virtual screening methodology in which ligands are fully and continuously flexible.

ICM allows one to draw a molecule directly in 3D with full undo/redo support and check its fit to a protein binding pocket. This environment is called a 3D ligand editor. The editor functionality is described in the User Manual.

The electrostatic potential can be projected on a molecular surface for the identification of possible binding sites.

[ Genomics | Dotplot | Alignment | Multi-alignment | Trees,clusters,evolution | Searches | Plotting ]

read sequence s_icmhome + "zincFing.seq" 
 plotSeqDotMatrix 2drp_d 3znf_m \ 
 "Two z-finger peptide" "Human Enhancer Domain" 5 20

(if the macro complains about s_psViewer , set it in Preferences/Directories , or reassign, e.g. s_psViewer = "display" ) Here the color shows the local significance of the alignment. You can change the method to calculate probability, color scheme and residue comparison matrices and calculate it interactively or in batch.

read sequence s_icmhome + "sh3.seq" 
 show Align(Fyn Spec)   # the probability will be shown

The ICM alignment functions and commands are summarized in the alignment section.

Read any number of sequences in fasta or swissprot formats and automatically align the sequences, interactively or in a batch. It will look like this:

# Consensus ...#.^.YD%..+~..-#~# K~-.#~##.~~..~WW.#.   ~~.~G%#P. 
Fyn     ----VTLFVALYDYEARTEDDLSFHKGEKFQILNSSEGDWWEARSLTTGETGYIPS 
Spec    DETGKELVLALYDYQEKSPREVTMKKGDILTLLNSTNKDWWKVE--VNDRQGFVP- 
Eps8    KTQPKKYAKSKYDFVARNSSELSM-KDDVLELILDDRRQWWKVR---NSGDGFVPN 
 
# nID 7 Lmin 56 ID 11.5 % 
#MATGAP gonnet 2.4 0.15

read sequence s_icmhome + "sh3.seq" 
 group sequences sh3 
 align sh3 
 show sh3

The ICM alignment functions and commands are summarized in the alignment section.

The Algorithm The multiple sequence alignment is produced using hierarchical clustering of the sequences based on sequence similarity calculated with the ZEGA alignment (a modification of the Needleman and Wunsch algorithm permitting zero gap-end penalties, ZEGA alignment) and Gonnet et al., residue substitution matrix [gon92]. The branches are then aligned in the tree order to form profiles that are progressively aligned to other profiles or sequences to form multiple sequence alignment.

Relationships between sequences can be presented in three forms: as evolutionary trees (ICM uses the neighbor-joining method for tree construction); as 2D distribution of sequences using the two main principal axes (use plot2Dseq macro); as 3D distribution. This can be analyzed in stereo using controls of molecular graphics (use ds3D macro: ds3D Distance(alig) Name(alig) ).

Take a matrix and represent it in 3D in a variety of forms. View it in stereo, color, label, transform with the mouse. Example: read matrix s_icmhome + "def" make grob def solid color display

• ICM-browser and ICM-browser-pro (distributed from a single package)
• ICM-chemist and ICM-chemist-pro (distributed from a single package)
• ICM-pro with options including bioinformatics, Poisson electrostatics, chemistry and cheminformatics, homology modeling, docking, virtual ligand screening (VLS).

• Windows
• Linux and Unix
• Macintosh

The modules have the following features: ICM-main

• shell for molecules, numbers, strings, vectors, matrices, tables, sequences, alignments, profiles, 3D maps, 3D graphical objects, 2D chemical tables/spreadsheets, images
• ICM-language and macros
• graphics, stereo
• imaging and vectorized postscript
• animation and movies
• mathematics, statistics, plotting
• presentation of the results in html format
• user-defined and automated interpretation of web links
• HTML-form-output interpretation
• pairwise and multiple sequence alignments, evolutionary trees, clustering
• secondary structure prediction and assignment, property profiles, pattern searching
• superpositions, structural alignment, Ramachandran plots
• protein quality check
• analytical molecular surface
• calculations of surface areas and volumes
• cavity analysis
• symmetry operations, access to 230 space groups
• database fragment search
• identification of common substructures in PDB
• read pdb, mol2, csd, build from sequence
• energy, solvation, MIMEL, side-chain entropies, soft van der Waals, tethers, distance and angular restraints
• local minimization
• ab initio peptide structure prediction by the Biased Probability Monte Carlo method
• loop simulations
• side-chain placement

• electrostatic free energy calculated by the boundary element method
• coloring molecular surface by electrostatic potential
• binding energy (electrostatic solvation component)
• maps of electrostatic potential and its isopotential contours

• indexing of chemical databases in SD, mol2 and csd format
• searching and extracting from the indexed databases
• fast grid potentials
• scripts for flexible ligand docking
• scripts for protein-protein docking
• 2D (SMILES) to 3D conversion, type and charge assignment, mmff geometry optimization, low-energy rotamer generation
• refinement in full atom representation

ICM chemistry (also, see here)

• scripting access to the internal chemical spreadsheets and external chemical databases in MySQL (via molcart) and SQL lite.
• Various operations on chemical tables (see below)
• integration with the docking engine and interactive ligand editing

Operations on chemical tables:

• Calculate various properties and descriptors from 2D chemical table
• Standardize chemical structure (change ambiguous depictions of some functional groups to a standard form according to user defined tables)
• Build QSAR type prediction models from a chemical spreadsheet and apply those models to new chemical tables
• Convert Smiles to 2D and the opposite operation
• Generate 2D Depiction from a 3D or 0D chemical
• Convert a 2D chemical to 3D
• Generate 3D Conformers
• Generate Tautomers
• Generate Stereoisomers, assign and manipulate stereo centers
• Align/Color By 2D Scaffold
• Cluster chemicals by either fingerprint similarity or external distance matrices
• Compare two chemical sets for common elements
• Sort a table and select duplicate rows in a table
• Create/Modify Markush objects
• Enumerate a combinatorial chemical library from scaffold and R-groups
• R-Group Decomposition of a chemical spreadsheet
• Enumerate a chemical library by reaction(s) and reactants
• Various forms of multiple chemical superposition (both 2D and 3D)

ICM-bioinformatics

• fast comparison and redundancy removal of millions of genomic or protein sequences
• multiple EST clustering, alignment and consensus derivation
• database indexing and manipulations
• functions to evaluate sequence-structure similarity
• scripts to recognize remote similarities in the protein sequence and PDB databases
• search a pattern through a database
• searching profiles and patterns from the Prosite database through a sequence
• HTML representation of the search results with interpretation of links
• interactive editor of sequence-structure alignment
• automated building of models by homology with loop sampling and side-chain placement (fast homology model building combined with the database loop search is a separate module which is ICM Homology).

• sequence-structure alignment (threading)
• ultra-fast automated homology model building with a database loop search
• loop modeling and refinement, side-chain placement
• surface analysis