Software tools for structure elucidation of synthetic and natural product peptide mixtures by LC-IM-MS

Peptides are biological molecules that are composed of chains of amino acid building blocks linked by chemical bonds. Peptides have many crucial roles in biology, including as defensive molecules, hormones, and allowing the immune system to recognise infected cells. Accumulation of clusters of peptides in the brain is implicated in Alzheimer's disease. Therefore peptides are commonly employed in research aimed at understanding basic biology and mechanisms of disease, as well as acting as starting points for new therapeutics. Peptide drugs have already been approved for conditions ranging from cancer to constipation and the number of peptides coming to market has accelerated in recent years.

Peptides can be produced for experiments and therapies through natural fermentation of appropriate organisms, although it is often more convenient to produce them chemically from amino acids in the laboratory. Both routes afford material that contains not just the desired peptide but typically also other molecules which may have very similar chemical structures to the one of interest. When making peptides chemically, the structures of many of these other by-products can be predicted from knowledge of the reactions and side-reactions taking place. Although purification procedures help to remove many impurities, this adds time and expense, wastes sample and is never 100% effective. Unfortunately, the peptide impurities may themselves have very intense biological activity which can cause confusion and misunderstanding if experimenters are unaware of their presence. Careful analysis of peptide samples is therefore highly desirable to determine exactly what is present, and indeed is a required part of the quality control for peptides destined to be used as drugs.

Peptide sample analysis on modern instruments involves several stages. In the first stage, peptides in an aqueous solution are separated as well as possible according to their physical properties. This solution then passes through a high voltage needle which leads to a spray from which the solvent evaporates to leave a gas containing the peptide molecules with an electrical charge. These charged peptides are then sorted by their size and electric charge, before undergoing a final step in which they are broken into pieces and their masses are measured. As each sample may contain many different peptides, these experiments produce huge quantities of data and the challenge is to relate this to the peptide structures in terms of the types of amino acid that they contain and the order in which these amino acids are connected together. The goal of our proposal is to write new computer software to simplify this task by automatically labelling the data with the identities of the peptides that can be detected. To do this, the software will simulate the synthesis of peptides and their by-products, make predictions about their analytical properties and compare these predictions with the experimental data. The software will run on typical laboratory and office computers and will feature a friendly graphical user interface. It will have a modular design that will allow us to add and refine features in the future in response to the demands of industry and academia. We will initially make a core set of features available for free, but subsequently offer enhanced versions catering for commercial users for a fee that will support future development.

The objective of the proposal is to develop a software system for automated structural annotation of liquid chromatography-ion mobility-mass spectrometry datasets of peptides originating from non-proteomics workflows, and in particular chemical peptide synthesis. The software will have a modular object oriented architecture and will feature objects with the purpose of (i) prediction of structures of peptide products and by-products of a synthesis; (ii) prediction of reverse phase HPLC retention times of these compounds; (iii) estimation of their gas phase ion mobility; (iv) prediction of collision induced dissociation (CID) and electron transfer dissociation (ETD) mass spectra; (v) scoring of structures against experimental data and annotation. Peptides will be represented as labelled graphs to capture the possible structural outcomes of (bio)synthetic reaction pathways. HPLC retention and ion mobility will be predicted using computationally rapid parameterised approaches based on algorithms previously reported in the context of proteomics. The principal CID and ETD fragmentation pathways (a, b/y, c/z ions) will be considered for calculation of mass spectra. MS/MS spectra will be used as the primary scoring criteria for candidate peptide structures, with predicted LC retention and ion mobility used as secondary criteria to help resolve ambiguous assignments and to flag up topological variations such as cyclic structures. The core functions of the software will be written in platform independent code which will be wrapped with a user interface for Windows.

The output from our proposed project will be a new software package for structural characterisation of components of peptide mixtures using modern liquid chromatography-ion mobility-mass spectrometry instruments. The existing software market is heavily dominated by tools for proteomics applications and we will instead address a need for impurity profiling of synthetic peptides while also keeping an eye towards applications in peptide natural products analysis. There are numerous commercial custom peptide suppliers and the level of characterisation supplied with their products is typically minimal and likely insufficient for many end users who may nevertheless take such products to be adequately pure on trust. Tools that assist suppliers to optimise their manufacturing processes and provide a more thorough quality analysis for their customers will be of benefit to all parties. The downstream users of peptides who will benefit are involved in fields of research such as immunology that have a direct bearing on human health through the development of new therapeutics and vaccines. Peptide therapeutics for diverse conditions are being developed at an increasing rate - more than 50 compounds have been approved and the market is worth billions. Long terms impacts (> 10 y) of our work will therefore be felt through improvements in patient care and the contribution of the pharmaceutical and biotechnology industries to the UK economy.

We will publicise our new tools to academia and industry through open access publications which will be deposited in repositories including Europe PubMed Central and the Cardiff University institutional repository. Use of electronic supplementary information will make code, parameters and example datasets available. We will also make use of the institutional website and community websites for publicising our work. Over the course of the grant we will present at major peptide and mass spectrometry conferences, including a Gordon Research Conference and a British Mass Spectrometry Society meeting. We will communicate our work with the public and school pupils through participation in outreach events under the auspices of Cardiff University and the Royal Society of Chemistry. We will also develop a workshop for A-level students to illustrate the power of modern mass spectrometry in analysis of biomolecules.

We intend to commercialise future versions of the software with tiered license fees for academic and commercial users. We will ensure the software is user friendly and runs under popular operating systems (initially Windows) on commodity hardware. A range of options aimed at being attractive for industrial users will be developed and licensed at extra cost. Cardiff University Research, Innovation and Enterprise Services will assist with commercialisation by advising on intellectual property protection and administering licenses and fees through University College Cardiff Consultants (UC3) Ltd. Revenues from license sales will be used to fund maintenance releases and enhancement of the software.

The proposed work will give a PDRA experience and training in scientific software development, user interface design and the principles of peptide chemistry and mass spectrometry. The skills acquired are in demand from both the life-science and analytical instrument industries but it is expected that the destination of the PDRA could also be other sectors employing skilled programmers such as financial services and digital creative industries (games, on-line media etc.). The work will also feed into student training at MSc and undergraduate level through final year and summer research projects in peptide synthesis and mass spectrometry. These activities will add to the UK skills base, boosting the knowledge economy and ensuring the UK remains an attractive base for science and technology businesses.