I recently wanted to explore areas of machine learning that I do not usually interact with as part of my DPhil research on antibody drug discovery. This post explores how to train and deploy a speech generation model for parliamentary speeches in the style of Jeremy Corbyn and Boris Johnson. You can play around with the resulting model at https://con-schneider.github.io/theytalktoyou.html.
Continue reading
Last week I presented the above paper at group meeting. While a little different from a typical OPIG journal club paper, the data we have access to almost certainly suffers from the same range of (possible) biases explored in this paper.
Continue readingThis is an interesting paper by Zhavoronkov, et al. that recently got published in Nature Biotechnology as a brief communication: https://www.nature.com/articles/s41587-019-0224-x. The paper describes a new deep generative model called generative tensorial reinforcement learning (GENTRL), which enables optimization for synthetic feasibility, novelty, and biological activity. In this work, authors have deigned, synthesized, and experimentally validated molecules targeting discoidin domain receptor 1 (DDR1) in less than two months. The code for GENTRL is available here: https://github.com/insilicomedicine/gentrl.
Reference: Zhavoronkov, A. et al. Deep learning enables rapid identification of potent DDR1 kinase inhibitors. Nature Biotechnology 2019, 37, 1038-1040.
Manually tuning hyperparameters in a neural network is slow and boring. Using Bayesian Optimisation to do it for you is slightly less slower and you can go do other things whilst it’s running. Susan recently highlighted some of the resources available to get to grips with GPyOpt. Below is a copy of a Jupyter Notebook where we walk through a couple of simple examples and hopefully shed a little bit of light on how the algorithm works.
Continue readingNeurIPS is the largest machine learning conference (by number of participants), with over 8,000 in 2017. This year, the conference will be held in Vancouver, Canada from 8th-14th December.
Recently, the list of accepted papers was announced, with 1430 papers accepted. Here, I will highlight several of potential interest to the chem-/bio-informatics communities. Given the large number of papers, these were selected either by “accident” (i.e. I stumbled across them in one way or another) or through a basic search (e.g. Ctrl+f “molecule”).
Continue readingHi everyone,
My blogpost this time around is a list of conferences popular with OPIGlets. You are highly likely to see at least one of us attending or presenting at these meetings! I’ve tried to make it as exhaustive as possible (with thanks to Fergus Imrie!), listing conferences in upcoming chronological order.
(Most descriptions are slightly modified snippets taken from the official websites.)
In this blog post I will share with you the materials that I found most useful when I started doing some Bayesian Optimisation in my research. Bear in mind, I am a Chemist by training, so I approached this topic from a non-mathematical background (my eyes have to be persuaded to look at mathematical equations). Out of all the materials I have come across, I found these to be the most accessible.
Continue readingThe kernel trick is a well known method in machine learning for producing a real-valued measure of similarity between data points in any number of settings. Kernel methods for network analysis provide a way of assigning real values to vertices of the graph. These values may correspond to similarity across any number of graphical properties such as the neighbours they share, or more dynamic context, the influence that change in the state of one vertex might have on another.
By using the kernel trick it is possible to approximate the distribution of features on the vertices of a graph in a way that respects the graphical relationships between vertices. Kernel based methods have long been used, for instance in inferring protein function from other proteins within Protein Interaction Networks (PINs).
Continue ReadingIn cheminformatics, there are many possible ways to encode chemical data represented by small molecules and proteins, such as SMILES, fingerprints, chemical descriptors etc. Recently, utilising graph-based methods for machine learning have become more prominent. In this post, we will explore why representing molecules as graphs is a natural and suitable encoding. Continue reading
Generating low-energy molecular conformers is important for many areas of computational chemistry, molecular modeling and cheminformatics. Many tools have been developed to generate conformers, including BALLOON (1), Confab (2), FROG2 (3), MOE (4), OMEGA (5) and RDKit (6). The search algorithm implemented in these tools can be broadly classified as either systematic or stochastic. These algorithms primarily focus on generating geometrically diverse low-energy conformers. Here, we are interested in finding lowest energy conformation of a molecule instead of achieving geometric diversity and Bayesian optimization is used to find the lowest energy conformation (7). Continue reading