The Robot Scientist Adam.
We review the main components of autonomous scientific discovery, and how they lead to the concept of a Robot Scientist. This is a system which uses techniques from artificial intelligence to automate all aspects of the scientific discovery process: it generates hypotheses from a computer model of the domain, designs experiments to test these hypotheses, runs the physical experiments using robotic systems, analyses and interprets the resulting data, and repeats the cycle. We describe our two prototype Robot Scientists: Adam and Eve. Adam has recently proven the potential of such systems by identifying twelve genes responsible for catalysing specific reactions in the metabolic pathways of the yeast Saccharomyces cerevisiae. This work has been formally recorded in great detail using logic. We argue that the reporting of science needs to become fully formalised and that Robot Scientists can help achieve this. This will make scientific information more reproducible and reusable, and promote the integration of computers in scientific reasoning. We believe the greater automation of both the physical and intellectual aspects of scientific investigations to be essential to the future of science. Greater automation improves the accuracy and reliability of experiments, increases the pace of discovery and, in common with conventional laboratory automation, removes tedious and repetitive tasks from the human scientist.
The EXACT description of biomedical protocols.%
This paper presents the design, implementation and performance of a reliable multicast transport protocol called RMTP. RMTP is based on a hierarchical structure in which receivers are grouped into local regions or domains and in each domain there is a special receiver called a Designated Receiver (DR) which is responsible for sending acknowledgments periodically to the sender, for processing acknowledgements from receivers in its domain and for retransmitting lost packets to the corresponding receivers. Since lost packets are recovered by local retransmissions as opposed to retransmissions from the original sender, end-to-end latency is significantly reduced, and the overall throughput is improved as well. Also, since only the DRs send their acknowledgments to the sender, instead of all receivers sending their acknowledgments to the sender, a single acknowledgement is generated per local region, and this prevents acknowledgement implosion. Receivers in RMTP send their acknowledgments to the DRs periodically, thereby simplifying error recovery. In addition, lost packets are recovered by selective repeat retransmissions, leading to improved throughput at the cost of minimal additional buffering at the receivers. This paper also describes the implementation of RMTP and its performance on the Internet.
The EXACT description of biomedical protocols.%
This paper presents the design, implementation and performance of a reliable multicast transport protocol called RMTP. RMTP is based on a hierarchical structure in which receivers are grouped into local regions or domains and in each domain there is a special receiver called a Designated Receiver (DR) which is responsible for sending acknowledgments periodically to the sender, for processing acknowledgements from receivers in its domain and for retransmitting lost packets to the corresponding receivers. Since lost packets are recovered by local retransmissions as opposed to retransmissions from the original sender, end-to-end latency is significantly reduced, and the overall throughput is improved as well. Also, since only the DRs send their acknowledgments to the sender, instead of all receivers sending their acknowledgments to the sender, a single acknowledgement is generated per local region, and this prevents acknowledgement implosion. Receivers in RMTP send their acknowledgments to the DRs periodically, thereby simplifying error recovery. In addition, lost packets are recovered by selective repeat retransmissions, leading to improved throughput at the cost of minimal additional buffering at the receivers. This paper also describes the implementation of RMTP and its performance on the Internet.
Logic and the Automatic Acquisition of Scientific Knowledge An Application to Functional Genomics.

Locational distribution of gene functional classes in Arabidopsis thaliana.

Decision Trees for Hierarchical Multilabel Classification A Case Study in Functional Genomics.

An ontology for a Robot Scientist.
doi:10.1093/bioinformatics/btl207
Decision Trees for Hierarchical Multilabel Classification A Case Study in Functional Genomics.

Functional bioinformatics for Arabidopsis thaliana.

Functional bioinformatics for Arabidopsis thaliana.

The Robot Scientist Project.
doi:10.1093/bioinformatics/btl207
Hierarchical Multi%2Dclassification with Predictive Clustering Trees in Functional Genomics.
This paper investigates how predictive clustering trees can  be used to predict gene function in the genome of the yeast Saccharomyces   cerevisiae. We consider the MIPS FunCat classification scheme,  in which each gene is annotated with one or more classes selected from  a given functional class hierarchy. This setting presents two important  challenges to machine learning: (1) each instance is labeled with a set of  classes instead of just one class, and (2) the classes are structured in a hierarchy;  ideally the learning algorithm should also take this hierarchical  information into account. Predictive clustering trees generalize decision  trees and can be applied to a wide range of prediction tasks by plugging  in a suitable distance metric. We define an appropriate distance metric  for hierarchical multi-classification and present experiments evaluating  this approach on a number of data sets that are available for yeast.
Scalable Multi%2DRelational Association Mining.

Confirmation of data mining based predictions of protein function.

Predicting gene function in Saccharomyces cerevisiae.
Motivation: S.cerevisiae is one of the most important model organisms, and has has been the focus of over a century of study. In spite of these efforts, 40 % of its open reading frames (ORFs) remain classified as having unknown function (MIPS: Munich Information Center for Protein Sequences). We wished to make predictions for the function of these ORFs using data mining, as we have previously successfully done for the genomes of M.tuberculosis and E.coli. Applying this approach to the larger and eukaryotic S.cerevisiae genome involves modifying the machine learning and data mining algorithms, as this is a larger organism with more data available, and a more challenging functional classification. Results: Novel extensions to the machine learning and data mining algorithms have been devised in order to deal with the challenges. Accurate rules have been learned and predictions have been made for many of the ORFs whose function is currently unknown. The rules are informative, agree with known biology and allow for scientific discovery. Availability: All predictions are freely available from
Data Mining the Yeast Genome in a Lazy Functional Language.
mining the yeast genome in a lazy
Machine learning of functional class from phenotype data.
Motivation: Mutant phenotype growth experiments are an important novel source of functional genomics data which have received little attention in bioinformatics. We applied supervised machine learning to the problem of using phenotype data to predict the functional class of ORFs in S. cerevisiae. Three sources of data were used: TRIPLES, EUROFAN and MIPS. The analysis of the data presented a number of challenges to machine learning: multi-class labels, a large number of sparsely populated classes, the need to learn a set of accurate rules (not a complete classi- cation), and a very large amount of missing values. We modi ed the algorithm C4.5 to deal with these problems.
How well do we understand the clusters found in microarray data%3F

Knowledge Discovery in Multi%2Dlabel Phenotype Data.
The biological sciences are undergoing an explosion in the amount of available data. New data analysis methods are needed to deal with the data. We present work using KDD to analyse data from mutant phenotype growth experiments with the yeast S. cerevisiae to predict novel gene functions. The analysis of the data presented a number of challenges: multi-class labels, a large number of sparsely populated classes, the need to learn a set of accurate rules (not a complete classification), and a very large amount of missing values. We developed resampling strategies and modified the algorithm C4.5 to deal with these problems.
Knowledge Discovery in Multi%2Dlabel Phenotype Data.
The biological sciences are undergoing an explosion in the amount of available data. New data analysis methods are needed to deal with the data. We present work using KDD to analyse data from mutant phenotype growth experiments with the yeast S. cerevisiae to predict novel gene functions. The analysis of the data presented a number of challenges: multi-class labels, a large number of sparsely populated classes, the need to learn a set of accurate rules (not a complete classification), and a very large amount of missing values. We developed resampling strategies and modified the algorithm C4.5 to deal with these problems.
The utility of different representations of protein sequence for predicting functional class.
Motivation: Data Mining Prediction (DMP) is a novel approach to predict protein functional class from sequence. DMP works even in the absence of a homologous protein of known function. We investigate the utility of different ways of representing protein sequence in DMP (residue frequencies, phylogeny, predicted structure) using the E. coli genome as a model. Results: Using the different representations DMP learnt prediction rules that were more accurate than default at every level of function using every type of representation. The most effective way to represent sequence was using phylogeny (75% accuracy and 13% coverage of unassigned ORFs at the most general level of function: 69% accuracy and 7% coverage at the most detailed). We tested different methods for combining predictions from the different types of representation. These improved both the accuracy and coverage of predictions, e.g. 40% of all unassigned ORFs could be predicted at an estimated accuracy of 60%, and 5% of unass...
The utility of different representations of protein sequence for predicting functional class.
Motivation: Data Mining Prediction (DMP) is a novel approach to predict protein functional class from sequence. DMP works even in the absence of a homologous protein of known function. We investigate the utility of different ways of representing protein sequence in DMP (residue frequencies, phylogeny, predicted structure) using the E. coli genome as a model. Results: Using the different representations DMP learnt prediction rules that were more accurate than default at every level of function using every type of representation. The most effective way to represent sequence was using phylogeny (75% accuracy and 13% coverage of unassigned ORFs at the most general level of function: 69% accuracy and 7% coverage at the most detailed). We tested different methods for combining predictions from the different types of representation. These improved both the accuracy and coverage of predictions, e.g. 40% of all unassigned ORFs could be predicted at an estimated accuracy of 60%, and 5% of unass...
Genome scale prediction of protein functional class from sequence using data mining.

