Christopher H. Bryant

Functional genomic hypothesis generation and experimentation by a robot scientist

The question of whether it is possible to automate the scientific process is of both great theoretical interest and increasing practical importance because, in many scientific areas, data are being generated much faster than they can be effectively analysed. We describe a physically implemented robotic system that applies techniques from artificial intelligence to carry out cycles of scientific experimentation. The system automatically originates hypotheses to explain observations, devises experiments to test these hypotheses, physically runs the experiments using a laboratory robot, interprets the results to falsify hypotheses inconsistent with the data, and then repeats the cycle. Here we apply the system to the determination of gene function using deletion mutants of yeast (Saccharomyces cerevisiae) and auxotrophic growth experiments. We built and tested a detailed logical model (involving genes, proteins and metabolites) of the aromatic amino acid synthesis pathway. In biological experiments that automatically reconstruct parts of this model, we show that an intelligent experiment selection strategy is competitive with human performance and significantly outperforms, with a cost decrease of 3-fold and 100-fold (respectively), both cheapest and random-experiment selection.

Are Grammatical Representations Useful for Learning from Biological Sequence Data?— A Case Study

This paper investigates whether Chomsky-like grammar representations are useful for learning cost-effective, comprehensible predictors of members of biological sequence families. The Inductive Logic Programming (ILP) Bayesian approach to learning from positive examples is used to generate a grammar for recognising a class of proteins known as human neuropeptide precursors (NPPs). Collectively, five of the co-authors of this paper, have extensive expertise on NPPs and general bioinformatics methods. Their motivation for generating a NPP grammar was that none of the existing bioinformatics methods could provide sufficient cost-savings during the search for new NPPs. Prior to this project experienced specialists at SmithKline Beecham had tried for many months to hand-code such a grammar but without success. Our best predictor makes the search for novel NPPs more than 100 times more efficient than randomly selecting proteins for synthesis and testing them for biological activity. As far as these authors are aware, this is both the first biological grammar learnt using ILP and the first real-world scientific application of the ILP Bayesian approach to learning from positive examples. A group of features is derived from this grammar. Other groups of features of NPPs are derived using other learning strategies. Amalgams of these groups are formed. A recognition model is generated for each amalgam using C4.5 and C4.5rules and its performance is measured using both predictive accuracy and a new cost function, Relative Advantage (RA). The highest RA was achieved by a model which includes grammar-derived features. This RA is significantly higher than the best RA achieved without the use of the grammar-derived features. Predictive accuracy is not a good measure of performance for this domain because it does not discriminate well between NPP recognition models: despite covering varying numbers of (the rare) positives, all the models are awarded a similar (high) score by predictive accuracy because they all exclude most of the abundant negatives.

A review of expert systems for chromatography

Abstract Expert systems for chromatography are reviewed. A taxonomy is proposed that allows present (and future) expert systems in this area to be classified and facilitates an understanding of their inter-relationship. All the systems are described focusing on the reasons for their development, what their purpose was and how they were to be used. The engineering methods, knowledge representations, tools and architectures used for the systems are compared and contrasted is a discussion covering all the stages of the development life cycle of expert systems. The review reveals that too often of expert systems for chromatography do not justify their decisions on engineering matters and that the literature suggests that many ideas advocated by knowledge engineeers are not being used.

Comparing the performance of object and object relational database systems on objects of varying complexity

This is the first published work to compare the performance of object and object relational database systems based on the objects complexity. The findings of this research show that the performance of object and object relational database systems are related to the complexity of the object in use. Object relational databases have better performance compared to object databases for fundamental database operations, with the exception of insert operations, on objects with low and medium complexity. For objects with high complexity, the object relational databases have better performance for update and delete operations.

Knowledge discovery in databases: application to chromatography

Abstract This article reviews emerging computer techniques for discovering knowledge from databases and their application to various sets of separation data. The data-sets include the separation of a diverse range of analytes using either liquid, gas or ion chromatography. The main conclusion is that the new techniques should help to close the gap between the rate at which chromatographic data is gathered and stored electronically and the rate at which it can be analysed and understood.

Towards an expert system for enantioseparations: induction of rules using machine learning

Abstract A commercially available machine induction tool was used in an attempt to automate the acquisition of the knowledge needed for an expert system for enantioseparations by high performance liquid chromatography using Pirkle-type chiral stationary phases (CSPs). Various rule-sets were induced that recommended particular CSP chiral selectors based on the structural features of an enantiomer pair. The results suggest that the accuracy of the optimal rule-set is 63% ± 3%, which is more than ten times greater than the accuracy that would have resulted from a random choice.

Inferring the Function of Genes from Synthetic Lethal Mutations

Techniques for detecting synthetic lethal mutations in double gene deletion experiments are emerging as powerful tool for analysing genes in parallel or overlapping pathways with a shared function. This paper introduces a logic-based approach that uses synthetic lethal mutations for mapping genes of unknown function to enzymes in a known metabolic network. We show how such mappings can be automatically computed by a logical learning system called extended hybrid abductive inductive learning (XHAIL).

A First Step towards Learning which uORFs Regulate Gene Expression

We have taken a first step towards learning which upstream Open Reading Frames (uORFs) regulate gene expression (i.e., which uORFs are functional) in the yeast Saccharomyces cerevisiae. We do this by integrating data from several resources and combining a bioinformatics tool, ORF Finder, with a machine learning technique, inductive logic programming (ILP). Here, we report the challenge of using ILP as part of this integrative system, in order to automatically generate a model that identifies functional uORFs. Our method makes searching for novel functional uORFs more efficient than random sampling. An attempt has been made to predict novel functional uORFs using our method. Some preliminary evidence that our model may be biologically meaningful is presented.

Data Mining via ILP: The Application of Progol to a Database of Enantioseparations

As far as this author is aware, this is the first paper to describe the application of Progol to enantioseparations. A scheme is proposed for data mining a relational database of published enantioseparations using Progol. The application of the scheme is described and a preliminary assessment of the usefulness of the resulting generalisations is made using their accuracy, size, ease of interpretation and chemical justification.

An ILP Refinement Operator for Biological Grammar Learning

We are interested in using Inductive Logic Programming ( ILP ) to infer grammars representing sets of biological sequences. We call these biological grammars. ILP systems are well suited to this task in the sense that biological grammars have been represented as logic programs using the Definite Clause Grammar or the String Variable Grammar formalisms. However, the speed at which ILP systems can generate biological grammars has been shown to be a bottleneck. This paper presents a novel refinement operator implementation, specialised to infer biological grammars with ILP techniques. This implementation is shown to significantly speed-up inference times compared to the use of the classical refinement operator: time gains larger than 5-fold were observed in