chemForma

We would like to create a Chemical Formula Specification for Mass Spectrometry that standardizes how chemical compositions are written.

Places this would be used

1. UniProt's Controlled vocabulary of posttranslational modifications (ptmlist.txt)
2. Unimod
3. PSI-MOD
4. ProForma

Goals of this project

1. Create a unified specification for a chemical formula to be written as a single line of ASCII text
2. Human readable, machine parsable, and optimized for the simple and common cases (e.g. H2O)
3. Handle condensed formulas
4. Handle specific MS use cases
   • Glycan atom support
   • Isotope support
   • Negative cardinalities

Do we need that? Isn't this already settled?

Given the goals above, it seems like something should already exist for this. But in my searching, I've found only the following:

• Hill Convention

  The well known Hill Convention introduced using element symbols, cardinality, and specific ordering to chemical formulas. However, for MS, we desire additional features.

• InChI

  InChI has a required base layer for a chemical formula, but it simply uses Hill convention and doesn't go any further.

• SMILES

  SMILES is another common format, but only encodes structure and doesn't have a place for chemical formula. There are other extensions of SMILES (e.g. SMARTS) that seem to have the same general problems.

What do MS modification repositories do?

• UniProt (ptmlist.txt)

  UniProt maintains a list of modifications used in its protein entries called ptmlist.txt. This document encodes a correction formula and appears to simply be Hill Convention with additional support for negative cardinalities and extra spaces between the elements.

  H2 O

  H-7 N-1 O1

• Unimod

  Unimod lists a database field called Composition that is the delta between modified and unmodified. This format handles glycans and certain isotopes as named 'atoms', yet requires all cardinalities be wrapped in paranthesis.

  H(2) O

  H(-7) C(-3) N(-1) S(-1)

  C(-2) 13C(2) N(-2) 15N(2)

  Ac Hex HexNAc NeuAc(2)

• PSI-MOD (and RESID)

  After a survey of terms, PSI-MOD appears use Hill Convention, while handling isotopes using a prefix wrapped in paranthesis. There was no inherant glycan support as full elemental formulas were given. Also, spaces were used extensively: they separate elements from each other as well as from cardinalities.

  H 2 O

  C 0 H 0 N 0 O 0 S -1 Se 1

  (12)C 8 (13)C 4 H 20 (14)N 1 (15)N 1 O 2

Our Proposal

Given that nothing accomplishes all the goals laid out above, we must simply accept an existing format or propose the creation of a new standard.

Adopt Unimod's Composition structure

The Unimod composition field is the only approach that covers enough of the requirements to consider adoption. We would need to create a specification document and get buy-in from all relevant parties.

--- OR ---

Create a new standard: chemForma

Making a new standard could make adoption more difficult, but it would allow input from the community and yield a (hopefully) better result. We would start with the following rules:

Rule 1

A formula will be composed of pairs of atoms and their corresponding cardinality. Pairs may be separated by spaces, but are not required to be. Also, Hill convention for ordering is perferred, but not required.

C12H20O2

Rule 2

Atoms will consist of all known periodic elements as well as known glycan residues (likely a subset from here, we will defer to glycan experts). If glycan symbols conflict with themselves or element symbols in such a way that ambiguities occur, we will consider requiring spaces between 'atoms' (see Rule #1).

Hex2Man

Rule 3

Cardinalities must be positive or negative integer values. Zero is not supported. If a cardinality is not included with an atom, it is assumed to be 1.

HN-1O2

Rule 4

Isotopes will be handled by prefixing the atom with its isotopic number in square brackets.

[13]C2[12]C-2H2N

Rule 5

A formula may include repeated (condensed) sections using parentheses and an extra cardinality. For example CH₃(CH₂)₄CH₃ would be written as:

CH3(CH2)4CH3