Skip to main content

An overview and invitation to contribute to ChEMBL curation with PPDMs

PPDMs has been in the making for more than a year and is a follow-up on a conference paper we published in 2012. As in 2012, our objective is to map small molecule binding sites to protein domains, the structural units that form recurring building blocks in the evolution of proteins. An application note describing PPDMs is just out in Bioinformatics.

Mapping small molecule binding to protein domains

The mapping facilitates the functional interpretation of small molecule-protein interactions - if you understand which domain in a protein is targeted, you are in a better position to anticipate the downstream effect.  Mapping small molecule binding to protein domains also provides a technical advantage to machine-learning approaches that incorporate protein sequence information as a descriptor to predict small molecule bioactivity. Reducing the sequence descriptor to the part that mediates small molecule binding increases the informative content of the descriptor. This is best exemplified by the domain-poisoning problem, illustrated below.
Result of a hypothetical query using as input the rat Tyrosine-protein phosphatase Syp (P35235) - and one of the hits, retrieved from a BLAST query against the ChEMBL target dictionary - the rat Tyrosine-protein kinase SYK (Q64725). The significant e-value for this query results from high scoring alignments of the SH2 domains. At the same time, the overlap between small molecules binding both proteins is expected to be low.

A simple heuristic

For individual experiments, it is often quite trivial to decide which domain was targeted. For example, medicinal chemists know whether their compound is a kinase inhibitor or one of a handful of SH2 inhibitors. This knowledge, while easily gleaned by the expert, is implicit and cannot be accessed programmatically. Hence we were motivated to implement a solution that could achieve this across as many measured bioactivities as possible.

Our initial implementation of mapping small molecules to protein domains consisted of a simple heuristic: Identify domains with known small molecule interaction and use these domains as a look-up when mapping measured bioactivities to protein domains. This process is illustrated in the figure below.

A catalogue of validated domains was extracted from assays against single-domain proteins (step 1, 2) and projected onto measured bioactivities in ChEMBL (step 3). Three possible outcomes are: i) A successful mapping if exactly one of the Pfam-A domain models from the catalogue matches the sequence; ii) No mapping if none of the Pfam-A domain models from the catalogue match the sequence; iii) A conflicting mapping if multiple domain models from the catalogue match the sequence.
Despite its simplicity, this method works surprisingly well, owing to the fact that protein domains that are relevant to drug discovery are prioritised in Pfam-A model curation. Another factor that contributes here is the conservative route taken by many drug discovery projects that focus on targets that are in well characterised protein families. However, as illustrated by the cases labelled ii) and iii), some constellations are not covered by the simple heuristic.

A public platform to review and improve mappings


Measured activities in ChEMBL falling into category iii) from the illustration above amount to only a fraction of the total but often reflect interesting biology. DHFR-TS for example is a multi-functional enzyme combining both a DHFR and Thymidylate_synt domain that occurs in the group of bikonts, which includes Trypanosoma and Plasmodium. In humans (and all metazoa), these domains occur as separate enzymes.
Small molecule inhibitors exist for both domains, DHFR (yellow, with Pyrimethamine) and Thymidylate synthase (blue, with Deoxyuridine monophosphate).
We built PPDMs as a platform to resolve such cases. PPDMs aggregates information that supports manual mapping assignments based on medicinal chemistry knowledge. New mappings can be  committed to the PPDMs logs and then transferred to the ChEMBL database in future releases.

The Conflicts section on the website summarises conflicts (cases that correspond to category iii as discussed above) that were encountered when the mapping was applied to measured activities in the ChEMBL database and offers an interface to resolve them.

The Evidence section provides the full catalogue of domains for which we found evidence of small molecule binding. Evidence for the majority of domains in this list is provided in the form of measured bioactivities in ChEMBL, while in a few cases we provide a reference to the literature. These are cases where well-known domains occur exclusively in multi-domain architectures, such as 7tm_2 and 7tm_3. The catalogue can be downloaded in full from this section.

PPDMs also provides logs of individual assignments - these can be queried by date, user and comments left when the assignment was made. A log of all assigned mappings can be downloaded from this section. Another way to review assigned mappings is through the Resolved section, where assignments are grouped by domain architecture.

We invite everyone with an interest in the matter to sign up with PPDMs, whether it's simply for playing around, resolving remaining conflicts, or reviewing existing assignments.  Please get in touch and we'll sort out a login for you!

felix

Comments

Popular posts from this blog

UniChem 2.0

UniChem new beta interface and web services We are excited to announce that our UniChem beta site will become the default one on the 11th of May. The new system will allow us to better maintain UniChem and to bring new functionality in a more sustainable way. The current interface and web services will still be reachable for a period of time at https://www.ebi.ac.uk/unichem/legacy . In addition to it, the most popular legacy REST endpoints will also remain implemented in the new web services: https://www.ebi.ac.uk/unichem/api/docs#/Legacy Some downtime is expected during the swap.  What's new? UniChem’s current API and web application is implemented with a framework version that’s not maintained and the cost of updating it surpasses the cost of rebuilding it. In order to improve stability, security, and support the implementation and fast delivery of new features, we have decided to revamp our user-facing systems using the latest version of widely used and maintained frameworks, i

A python client for accessing ChEMBL web services

Motivation The CheMBL Web Services provide simple reliable programmatic access to the data stored in ChEMBL database. RESTful API approaches are quite easy to master in most languages but still require writing a few lines of code. Additionally, it can be a challenging task to write a nontrivial application using REST without any examples. These factors were the motivation for us to write a small client library for accessing web services from Python. Why Python? We choose this language because Python has become extremely popular (and still growing in use) in scientific applications; there are several Open Source chemical toolkits available in this language, and so the wealth of ChEMBL resources and functionality of those toolkits can be easily combined. Moreover, Python is a very web-friendly language and we wanted to show how easy complex resource acquisition can be expressed in Python. Reinventing the wheel? There are already some libraries providing access to ChEMBL d

LSH-based similarity search in MongoDB is faster than postgres cartridge.

TL;DR: In his excellent blog post , Matt Swain described the implementation of compound similarity searches in MongoDB . Unfortunately, Matt's approach had suboptimal ( polynomial ) time complexity with respect to decreasing similarity thresholds, which renders unsuitable for production environments. In this article, we improve on the method by enhancing it with Locality Sensitive Hashing algorithm, which significantly reduces query time and outperforms RDKit PostgreSQL cartridge . myChEMBL 21 - NoSQL edition    Given that NoSQL technologies applied to computational chemistry and cheminformatics are gaining traction and popularity, we decided to include a taster in future myChEMBL releases. Two especially appealing technologies are Neo4j and MongoDB . The former is a graph database and the latter is a BSON document storage. We would like to provide IPython notebook -based tutorials explaining how to use this software to deal with common cheminformatics p

ChEMBL 30 released

  We are pleased to announce the release of ChEMBL 30. This version of the database, prepared on 22/02/2022 contains: 2,786,911 compound records 2,157,379 compounds (of which 2,136,187 have mol files) 19,286,751 activities 1,458,215 assays 14,855 targets 84,092 documents Data can be downloaded from the ChEMBL FTP site: https://ftp.ebi.ac.uk/pub/databases/chembl/ChEMBLdb/releases/chembl_30/ Please see ChEMBL_30 release notes for full details of all changes in this release:  https://ftp.ebi.ac.uk/pub/databases/chembl/ChEMBLdb/releases/chembl_30/chembl_30_release_notes.txt New Deposited Datasets EUbOPEN Chemogenomic Library (src_id = 55, ChEMBL Document ID CHEMBL4689842):   The EUbOPEN consortium is an Innovative Medicines Initiative (IMI) funded project to enable and unlock biology in the open. The aims of the project are to assemble an open access chemogenomic library comprising about 5,000 well annotated compounds covering roughly 1,000 different proteins, to synthesize at least

Multi-task neural network on ChEMBL with PyTorch 1.0 and RDKit

  Update: KNIME protocol with the model available thanks to Greg Landrum. Update: New code to train the model and ONNX exported trained models available in github . The use and application of multi-task neural networks is growing rapidly in cheminformatics and drug discovery. Examples can be found in the following publications: - Deep Learning as an Opportunity in VirtualScreening - Massively Multitask Networks for Drug Discovery - Beyond the hype: deep neural networks outperform established methods using a ChEMBL bioactivity benchmark set But what is a multi-task neural network? In short, it's a kind of neural network architecture that can optimise multiple classification/regression problems at the same time while taking advantage of their shared description. This blogpost gives a great overview of their architecture. All networks in references above implement the hard parameter sharing approach. So, having a set of activities relating targets and molecules we can tra