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 train a single neural network as a binary multi-label classifier that will output the probability of activity/inactivity for each of the targets (tasks) for a given query molecule.

PyTorch is one of the most popular open source AI libraries at present. It's getting a lot of traction in research environments, it's deeply integrated with the NumPy ecosystem and it also implements a dynamic graph approach making it easier to debug.

We have some interesting references, we have data in ChEMBL, we have PyTorch and RDKit... what are we waiting for?

First of all we'll need to extract the data from ChEMBL and format it for our purpose. The following notebook explains step by step how to do it. The output will be a H5 file that you can also download from here in case you want go directly to the network training phase.

Notebook to extract the data

Nice! We have the data, let's go then through the main notebook and train a model!

Notebook to train the model

This was a simple example. We hope you enjoyed it and will be inspired to experiment with deeper architectures, skipping connections, different learning rate strategies, more epochs, early stopping... and so on!

Notebooks also available in GitHub

Comments

ChEMBL 34 is out!

We are delighted to announce the release of ChEMBL 34, which includes a full update to drug and clinical candidate drug data. This version of the database, prepared on 28/03/2024 contains: 2,431,025 compounds (of which 2,409,270 have mol files) 3,106,257 compound records (non-unique compounds) 20,772,701 activities 1,644,390 assays 15,598 targets 89,892 documents Data can be downloaded from the ChEMBL FTP site: https://ftp.ebi.ac.uk/pub/databases/chembl/ChEMBLdb/releases/chembl_34/ Please see ChEMBL_34 release notes for full details of all changes in this release: https://ftp.ebi.ac.uk/pub/databases/chembl/ChEMBLdb/releases/chembl_34/chembl_34_release_notes.txt New Data Sources European Medicines Agency (src_id = 66): European Medicines Agency's data correspond to EMA drugs prior to 20 January 2023 (excluding vaccines). 71 out of the 882 newly added EMA drugs are only authorised by EMA, rather than from other regulatory bodies e.g.

New SureChEMBL announcement

(Generated with DALL-E 3 ∙ 30 October 2023 at 1:48 pm) We have some very exciting news to report: the new SureChEMBL is now available! Hooray! What is SureChEMBL, you may ask. Good question! In our portfolio of chemical biology services, alongside our established database of bioactivity data for drug-like molecules ChEMBL , our dictionary of annotated small molecule entities ChEBI , and our compound cross-referencing system UniChem , we also deliver a database of annotated patents! Almost 10 years ago , EMBL-EBI acquired the SureChem system of chemically annotated patents and made this freely accessible in the public domain as SureChEMBL. Since then, our team has continued to maintain and deliver SureChEMBL. However, this has become increasingly challenging due to the complexities of the underlying codebase. We were awarded a Wellcome Trust grant in 2021 to completely overhaul SureChEMBL, with a new UI, backend infrastructure, and new f

Accessing SureChEMBL data in bulk

It is the peak of the summer (at least in this hemisphere) and many of our readers/users will be on holiday, perhaps on an island enjoying the sea. Luckily, for the rest of us there is still the 'sea' of SureChEMBL data that awaits to be enjoyed and explored for hidden 'treasures' (let me know if I pushed this analogy too far). See here and here for a reminder of SureChEMBL is and what it does. This wealth of (big) data can be accessed via the SureChEMBL interface , where users can submit quite sophisticated and granular queries by combining: i) Lucene fields against full-text and bibliographic metadata and ii) advanced structure query features against the annotated compound corpus. Examples of such queries will be the topic of a future post. Once the search results are back, users can browse through and export the chemistry from the patent(s) of interest. In addition to this functionality, we've been receiving user requests for local (behind the

Mapping lists of IDs in ChEMBL

In order to facilitate the mapping of identifiers in ChEMBL, we have developed a new type of search in the ChEMBL Interface. Now, it is possible to enter a list of ChEMBL IDs and see a list of the corresponding entities. Here is an example: 1. Open the ChEMBL Interface , on the main search bar, click on 'Advanced Search': 2. Click on the 'Search by IDs' tab: 3. Select the source entity of the IDs and the destination entity that you want to map to: 4. Enter the identifiers, you can either paste them, or select a file to upload. When you paste IDs, by default it tries to detect the separator. You can also select from a list of separators to force a specific one: Alternatively, you can upload a file, the file can be compressed in GZIP and ZIP formats, this makes the transfer of the file to the ChEMBL servers faster. Examples of the files that can be uploaded to the search by IDs can be found here . 5. Click on the search button: 6. You will be redirected to a search resul

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

The ChEMBL-og

Search This Blog