Welcome to the John Mitchell research group
We are an Informatics and Computational Chemistry research group. Our group is based in the Purdie building on the North Haugh in St Andrews.
Research areas we are interested in are enzyme catalysis, protein-ligand interactions, molecular evolution and structural bioinformatics, computational toxicology, prediction of solubility and other molecular properties, and the classification of drugs used for doping in sport.
Our clustering algorithm PFClust can automatically find the best number of clusters for a dataset, as described in this Open Access paper. Luna De Ferrari's talk on Active and Guided Learning of Enzyme Function from MLSB 2012, Basel, is available as a video online. There are also available presentations describing our recent work on Computational Enzymology, Predicting the Mechanism of Phospholipidosis, RF-Score: A new scoring function for Protein-Ligand affinity prediction, The Chemistry of Protein Catalysis and In silico calculation of aqueous solubility. Our recent solubility prediction method utilising machine learning can be found here.
Thanks to our Sponsors
NewsThree prizes won at the RSC Younger Members Symposium (Birmingham - 23rd June 2014).
- Rosanna Alderson - Best speaker of the Education & Outreach session
- James McDonagh - Best speaker of the Physical & Analytical session
- Rachael Skyner - First prize poster presentation of the Physical & Analytical session
Is Experimental Data Quality the Limiting Factor in Predicting the Aqueous Solubility of Druglike Molecules? is published in Molecular Pharmaceutics, the result of a collaboration with Dr David Palmer at Strathclyde (11 June 2014).
The Natural History of Biocatalytic Mechanisms is published in the prestigious Open Access journal PLoS Computational Biology (29 May 2014).
From sequence to enzyme mechanism using multi-label machine learning is published Open Access in BMC Bioinformatics (19 May 2014).
Two more new Open Access publications: Uniting Cheminformatics and Chemical Theory to Predict the Intrinsic Aqueous Solubility of Crystalline Druglike Molecules, J. Chem. Inf. Model. (2014) and Machine learning methods in chemoinformatics, WIREs Comput. Mol. Sci. (2014)
Rosanna Alderson wins a Poster Prize at the MGMS Young Modellers' Forum at SOAS in London, joining a prestigious list of previous winners that includes our group's alumnus Florian Nigsch (29 Nov 2013).
Our clustering algorithm PFClust now has its own web page on this site (4 Sep 2013).
Our new solubility prediction scheme is avaliable for download on its own page on this site.
4273pi: Bioinformatics education on low cost ARM hardware is currently the number 1 most viewed article for the last 30 days in BMC Bioinformatics (26 Aug 2013).
Full "Laplacianised" posterior naive Bayesian algorithm is published Open Access in Journal of Cheminformatics (23 Aug 2013)
4273pi: Bioinformatics education on low cost ARM hardware is published Open Access in BMC Bioinformatics (12 Aug 2013).
In silico target predictions: defining a benchmarking dataset and comparison of performance of the multiclass Na´ve Bayes and Parzen-Rosenblatt Window is published online in the Journal of Chemical Information and Modeling (8 July 2013).
PFClust: a novel parameter free clustering algorithm is published Open Access in BMC Bioinformatics (3 July 2013).
Predicting the protein targets for athletic performance-enhancing substances is published Open Access in the Journal of Cheminformatics (25 June 2013).
Neetika Nath and James McDonagh's poster wins second prize at the ScotCHEM Computational Symposium 2013 (14 June 2013).
Active and guided learning of enzyme function, Luna De Ferrari's talk from MLSB 2012, Basel, is available as a video online (2 May 2013).
Rosanna Alderson wins the Innovation Poster Prize at the 3rd Annual PhD Symposium on Computational Biology and Innovation at University College Dublin (7 Dec 2012)!
Our review paper Enzyme Informatics is published in Current Topics in Medicinal Chemistry.
Our paper Hierarchical virtual screening for the discovery of new molecular scaffolds in antibacterial hit identification is published in Journal of the Royal Society Interface (29 Aug 2012).
The first St Andrews Bioinformatics Symposium was held on 22 August.
Congatulations to Richard on passing his PhD viva (10 Aug 2012).
First-Principles Calculation of the Intrinsic Aqueous Solubility of Crystalline Druglike Molecules is published online as an ASAP article in J. Chem. Theory Comput. (25 Jul 2012).
Congatulations to Luna on passing her PhD viva (24 Jul 2012).
See John's slides from the May 2012 conference "Computational Chemogenomics to understand System Biology & Computational Medicinal Chemistry" in Geneva (23 May 2012).
Is EC class predictable from reaction mechanism? is published in BMC Bioinformatics, an Open Access journal (24 Apr 2012).
Predicting the mechanism of phospholipidosis is published as an Open Access paper in the Journal of Cheminformatics (26 Jan 2012).
Read the story of our iGEM jamboree in Amsterdam on the SULSA website (26 Jan 2012).
MACiE version 3.0 is released.
Classifying Molecules Using a Sparse Probabilistic Kernel Binary Classifier is published in the Journal of Chemical Information and Modeling.
The review article Informatics, machine learning and computational medicinal chemistry has now been published in Future Medicinal Chemistry.
There's a new page describing our scoring function RF-Score. It's free of charge to all users, enjoy!
Why Are Some Properties More Difficult To Predict than Others? A Study of QSPR Models of Solubility, Melting Point, and Log P is currently listed in the Top 20 most cited articles in the Journal of Chemical Information and Modeling for the last three years (16 Nov 2010).
Predicting Intrinsic Aqueous Solubility by a Thermodynamic Cycle is currently listed in the Top 20 most cited articles in Molecular Pharmaceutics for the last three years (16 Nov 2010).
Predicting Phospholipidosis Using Machine Learning is published as an Open Access paper in Molecular Pharmaceutics under the ACS Author Choice scheme. We pay, you read.
The Protein Ligand Database (PLD) is available online as a spreadsheet.
You may be interested in our publications (e.g. scoring functions, odour prediction), or a compilation of resources that we use in our work (datasets containing structures and properties, list of EC-PDB-CATH correspondences, etc.).