HELM

HELM (Hierarchical Editing Language for Macromolecules) is a machine-readable notation for describing complex biomolecules such as peptides, oligonucleotides, antibody-drug conjugates, and the small-molecule linkers that join them. It was developed by the Pistoia Alliance, a non-profit consortium of pharmaceutical companies (Pfizer, AstraZeneca, GlaxoSmithKline, Novartis and others), with work beginning around 2008 and the specification published openly and freely from 2012. The current widely used revision is HELM2 (2.0). It arose because neither plain protein sequences nor atom-level chemical file formats such as Molfile or SDF conveniently capture large therapeutic molecules that mix natural residues, non-natural monomers, and chemical conjugation.

The data model is explicitly hierarchical, spanning four levels: complex polymer, simple polymer, monomer, and atom. A simple polymer is a linear sequence of monomers of one type (for example PEPTIDE, RNA/DNA, or CHEM for a small-molecule component); a complex polymer combines several simple polymers plus their inter-polymer bonds and annotations. Each monomer is a named building block whose atom-and-bond structure is stored in a monomer library, so a short symbol stands in for a full chemical graph, much as an identifier does. Crucially, monomers are defined by the author, so there is no fixed vocabulary: non-standard residues can be referenced by library ID or given inline as SMILES (SMILES), letting HELM cover chemistries that pure sequence formats cannot.

Syntactically, a HELM string is divided into sections separated by dollar signs: a polymer-definition section, a connection section, a hydrogen-bond section, and an extended-annotations section. Polymers are written as an ID followed by monomers in curly braces, e.g. PEPTIDE1{A.C.[dL].G}, with multi-character or non-standard monomers bracketed and separated by periods; multiple polymers are joined with a pipe. Connections name the two polymers, the monomer positions, and their attachment points (R-groups), such as PEPTIDE1,CHEM1,1:R1-1:R1, which also expresses cyclic peptides and branched conjugates. HELM2 adds repeating units, unknown or ambiguous sections, monomer/polymer probabilities, and mixtures; the annotation section may hold arbitrary JSON.

In the graph-data and format-conversion landscape HELM sits alongside other macromolecular and chemical notations: it is more structural than linear sequences yet more compact than atom-level formats, and it interconverts with SMILES and Molfile when a monomer library is available. Related biopolymer and chemistry notations in this catalog include PDB, CML, InChI, and monomer-oriented alternatives such as BigSMILES for polymers; toolkits like RDKit and Open Babel handle the underlying small-molecule structures. Strengths are compactness, readability at the sequence level, and the ability to represent conjugates and non-natural chemistry unambiguously. Limitations are real: HELM is meaningless without the referenced monomer library, so interoperability depends on shared or embedded definitions; the notation is niche, largely confined to pharmaceutical R&D and tools such as the HELM Editor; and expressing highly branched or atypical topologies can become verbose and error-prone.