SMILECAS Notations Descriptions

1. Introduction
SMILES is an acronym for Simplified Molecular Input Line Entry System. It is a chemical notation system used to represent a molecular structure by a linear string of symbols. The SMILES notation system was specifically designed for computer use by chemists. The encoding rules for SMILES can be learned quickly and easily by anyone with any type of chemistry background. The history of SMILES notation as a chemical language and the basic encoding rules for SMILES have been presented by David Weininger (J. Chem. Inf. Comput. Sci. 28(1): 31-6).

This on-line help outlines the basic rules used to formulate a SMILES notation for a chemical structure. The encoding rules outlined here document focus directly on the SRC software programs.

Learning to write a SMILES notation for most chemicals is not difficult. However, writing a SMILES notation for a complicated ring system can be tricky and time-consuming. The SMILECAS Database (available on-line with our LogKow demo and from SRC as an add-on product to our estimation software) is extremely helpful and time-efficient in obtaining SMILES notations. This database contains the SMILES notations for 103,000 compounds; all you need is the CAS (Chemical Abstract Service) Reigistry number.

2. Encoding Rules
A SMILES notation depicts a molecular structure as a two-dimensional picture as if drawn on a piece of paper. A two-dimensional drawing of a single chemical structure is possible in many different forms. That is, a single structure can be depicted correctly by many different drawings. In a similar manner, a single structure can be depicted correctly by many different SMILES notations. In fact, any modestly large structure has literally dozens of SMILES notations that will correctly depict the structure. Any one of the correct depictions is acceptable for computer interaction.

SMILES notations are comprised of atoms (designated by atomic symbols), bonds, parentheses (used to show branching), and numbers (used to designate ring opening and closing positions). With the exception of designating ring positions, numbers are not used in SMILES notation.

2.1. Atoms
Atoms are represented by their atomic symbols. For example:

C is carbon N is nitrogen S is sulfur F is fluorine

I is iodine P is phosphorus O is oxygen Cl is chlorine

Upper and lower case letters are important. All aliphatic atoms are entered in upper case. All aromatic atoms are entered in lower case. The possible aromatic atoms are carbon, oxygen, sulfur, selenium and nitrogen. Other potential aromatic atoms are not currently allowed by the SRC programs because the current estimation methods used in the programs can not evaluate them.

Atoms with two letter atomic symbols, such as chlorine or bromine, must have the first letter entered in upper case. In the case of chlorine or bromine, the second letter of the atomic symbol can be either upper or lower case. The "r" in bromine's symbol is usually entered in lower case. It is suggested that the "l" in chlorine's symbol be entered in upper case ("L") because it is possible to mis-identify a lower case "l" and the number one "1". Therefore, chlorine can be entered as either Cl or CL and bromine can be entered as either Br or BR.

With very rare exception (see section 3.3), the hydrogen atom is not included in a SMILES notation. Hydrogen attachments are determined by the program. This greatly simplifies a SMILES notation. For example:

Compound    Molecular Formula    SMILES Notation
---------   -----------------    ---------------
Ethylene    CH2=CH2              C=C
Propylene   CH2=CH-CH3           C=CC
2-Butene    CH3-CH=CH-CH3        CC=CC

2.2. Bonds
The four basic bonds in SMILES notation are single, double, triple, and aromatic bonds. Single bonds do not need to be shown and are usually omitted. A single can be designated with the hyphen symbol "-". For example, a correct SMILES notation for propane is C-C-C ;however, there is no advantage to entering the single bond. Therefore, it is not normally used (the SRC programs automatically remove any hyphens entered in a SMILES string).

The double bond is designated by the equal symbol "=" and is required to identify double bond. The following examples illustrate the double bond:

Compound    Molecular Formula    SMILES Notation
---------   -----------------    ---------------
Ethylene    CH2=CH2              C=C
Propylene   CH2=CH-CH3           C=CC
2-Butene    CH3-CH=CH-CH3        CC=CC

The triple bond is designated by the number symbol "#" and is required to identify a triple bond. The following examples illustrate the triple bond:

Compound         SMILES Notation
-------------    ---------------
Acetylene        C#C
Propyne          C#CC
Butyne           C#CCC
Acetonitrile     CC#N
Acrylonitrile    C=CC#N

The aromatic bond has no designation. It is explicitly implied by a "lower case letter" for carbon, nitrogen, oxygen, selenium and sulfur. For example, a typical SMILES notation for benzene is c1ccccc1 and a typical notation for pyridine is n1ccccc1. The use of the numbers as ring opening and closing positions is discussed in section 2.4.


2.3 Branches
Branches in molecular structures are designated by enclosures in parentheses. The examples of SMILES given in the lists above represent straight, linear compounds. When a structure contains a branch, the SMILES Notation of the structure requires that the branch be designated in enclosed parentheses. The figure below illustrates branching.

As previously noted, a single structure can have more than one valid SMILES notation. As an example, valid SMILES notations for the isobutyric acid structure (above figure) include the following:

CC(C)C(=O)O
C(C)(C)C(=O)O
OC(=O)C(C)C
O=C(O)C(C)C

A branch can not begin a SMILES notation. For example, (C)CCO is an invalid SMILES notation. A branch must immediately follow the atom to which it is connected. If an atom has more than one branch, the branches are coded as consecutive pairs of parentheses. The tert-butanol structure shown above is an example. The order of the parentheses is not important; for example, tert-butanol can be either CC(C)(O)C or CC(O)(C)C.

A branch can not immediately follow a double bond symbol "=" or a triple bond symbol "#"; it must immediately follow the atom. For example: C=(CC)C is invalid; if the double bond is connected to the carbon inside the parentheses, the SMILES should be C(=CC)C; if the double bond is connected to the final carbon, the SMILES should be C(CC)=C.

"Nested branches" or "branches-within-branches" are allowed (and frequent needed). The following figure illustrates nested branches.

Dozens of different, valid SMILES notations could be coded for the structure above. The notation could begin at any carbon in the structure. For example, if the notation begins at the center-most carbon, the SMILES notation could be: C(C=C)(CC)(C(C)C)(C(C)(C)C)

The SMILES interpreter used in the SRC programs does not allow two or more consecutive left-sided (starting) parentheses such as "((" to be used. An example would be: CC((CC))CC. The reason is: two left-sided parentheses are never needed to correctly represent any structure; their use promotes poorly coded SMILES notations. SMILES notations are usually easiest to comprehend when they have the fewest number of possible branches! Unnecessary branching can complicate a SMILES notation. For example, butane is best coded as: CCCC although, it is valid to code it as: C(C(C(C))).

2.4. Cyclic Structures
The most difficult aspect of writing SMILES notations is writing a correct SMILES notation for a complicated ring system! Writing SMILES notations for structures containing only one or two rings is fairly simple however. The following encoding rules apply to all cyclic structures:

(1) Cyclic structures require numbers to indicate where the ring starts and stops. The numbers 1 through 9 are used to indicate the starting and terminating atoms.

(2) The SAME number is used to indicate the starting and terminating atom for each ring. The starting and terminating atom must be connected to each other!

(3) Each number that is used (1, 2, 3, etc.) MUST appear twice and ONLY twice in the entire SMILES notation. This rule has an exception in the recent MS-Windows versions of the SRC programs. A SMILES such as c1ccccc1c1ccccc1 is allowed...the programs convert this to c2ccccc2c1ccccc1.

(4) Numbers are entered immediately following the atoms used to indicate the starting and terminating positions. For example, a number can not follow a branch as in: c1ccccc(Br)1; this notation for bromobenzene should written as c1ccccc1(Br) or c1ccccc1Br.

(5) A starting or terminating atom can be associated with two consecutive numbers. For example, naphthalene can be coded as: c12ccccc1cccc2 (see the example below). The "12" following the first carbon indicates that the first carbon is connected to both of the following numbered carbons. Three consecutive number are not currently allowed by the SRC programs.

Examples are the best way to understand SMILES notations for cyclic structures. Several examples are illustrated here. The following concept has been found useful for writing SMILES notations for ring systems: (a) select one ring from the entire structure and label the starting and terminating atoms with the number 1; (b) begin at the starting atom and "snake your way" (draw a free-hand line) through the cyclic structure so that the "snake" passes every ring member once and finishes at the terminating atom. Number each starting and terminating atom of each subsequent ring as it is passed by the "snake". For complicated structures, it may be quite a puzzle with many possible solutions. The key is to select an appropriate ring to start. Once the "snake" has been drawn, simply write the SMILES notation by starting at the initial atom and then follow the "snake". The "snake" in the examples below is the curved line that ends at the arrow head. The "snake" starts at the starting atom and ends at the terminating atom. Remember that aromatic atoms are entered in lower case.

The following examples illustrate ring systems where the rings are not connected to each other at two or more atoms (not fused):

In certain types of ring systems, it is impossible to draw the "snake" completely through all rings. In these situations, it is necessary to use "ring branching". The examples of benzene and acenaphthene below demonstrate ring branching; neither of these structures require it, but it is available. The strychnine structure example needs it; a SMILES can not otherwise be written.

2.5. Aromatic Conversions
The SMILES interpreter in the SRC programs will convert certain aliphatic rings to aromatic rings if aromaticity is detected. For example, the following conversions are made:

C1=CC=CC=C1 ----> c1ccccc1 (benzene)

N1C=CC=C1 ----> n1cccc1 (pyrrole)

O1C=CC=C1 ----> o1cccc1 (furan)

S1C=CC=C1 ----> s1cccc1 (thiofuran)

N1=CC=CC=C1 ----> n1ccccc1 (pyridine)

Other single ring and fused ring conversions are also made. See section 3 for additional information pertaining to aromatic conversions.

2.6. Aromatic Limitations
Certain valid aromatic structures are flagged as "Illegal Structures" by some of the SRC programs. This is because the estimation techniques used by the program can not evaluate that type of structure. An example is the compound azulene which is comprised of a fused 7-member ring and a 5-member ring. It is flagged as illegal because a 7-member aromatic was found. Currently, the SRC programs will not accept an aromatic ring with 7 or more ring-members. Most estimation methods have not been extended to include 7-member aromatic rings due to lack of data. Azulene (and similar structures), can be estimated by entering the "upper case" (aliphatic) SMILES (e.g. C1=CC=C2C=CC=C2C=C1).

2.7. Notations For Selected Fragments
Most users who are new to writing SMILES notations have trouble coding certain chemical fragments. The following list should be useful:

Fragment              SMILES             Example
---------------              ----------               -----------
Nitro                    N(=O)(=O)            CCN(=O)(=O) nitroethane
Nitrate                 ON(=O)(=O)          CON(=O)(=O) methyl nitrate
Nitrite                  ON(=O)                 CON=O methyl nitrite
Sulfonic acid        S(=O)(=O)O          CS(=O)(=O)O methyl sulfonic acid
Cyanide/Nitrile      C#N                     CC#N methyl cyanide
Azide                   N=N#N                 CN=N#N methyl azide
Azido as               N+=N-                 N#N

2.8. Metals
Metals are designated by the atomic symbol of the metal enclosed in square brackets. The current versions of the SRC programs can accept the following metals: