Use of Automatic Relevance Determination in QSAR Studies Using Bayesian Neural Networks

Burden, Frank R.; Ford, Martyn G.; Whitley, David C.; Winkler, David A.

doi:10.1021/ci000450a

Cited by 123 publications

(79 citation statements)

References 40 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Like evolution, one or more methods will be found to be superior and will grow in strength. Indeed, consensus methods and Bayesian networks [83] already seem to be demonstrating reasonable results within good statistical practices. The general feeling among scientists in this field is that Bayesian networks offer considerable advantages.…”

Section: Discussionmentioning

confidence: 94%

Neural Networks in 3D QSAR

Livingstone

Manallack²

2003

QSAR Comb. Sci.

View full text Add to dashboard Cite

Abstract3D QSAR data analysis typically deals with large numbers of descriptors and various methods have been used to cope with this multivariate problem. One popular method, CoMFA, makes use of Partial Least Squares (PLS) regression. Alternative methods of data analysis to PLS have been explored including artificial neural networks (ANN). Within 3D QSAR, ANNs have been successful in producing models showing reasonable predictive capabilities. This review will examine recent studies in 3D QSAR and QSPR using ANN methods.

show abstract

Section: Discussionmentioning

confidence: 94%

Neural Networks in 3D QSAR

Livingstone

Manallack²

2003

QSAR Comb. Sci.

View full text Add to dashboard Cite

show abstract

“…The mathematics of Bayesian regularization is challenging and is not repeated here as it is described in numerous publications. [7][8][9][11][12][13][14][15][44][45][46][47] …”

Section: Sparse Learning Methodsmentioning

confidence: 99%

“…Our team has developed effective computational design and modelling techniques over the past 20 years, [7][8][9][10][11][12][13][14][15] and we have employed them to generate substantial scientific and commercial impact, culminating in more than 20 patents. These methods have been used to design and optimise green pesticides [16][17][18][19][20] in collaboration with Du Pont and Schering Plough, have discovered new peptides and small molecules to control stem cells and cancers, [21][22][23][24][25][26][27] are accelerating the development of biomaterials for implantation and stem cell culture, [28][29][30][31][32][33] have provided new scientific insight into the potential adverse properties of nanomaterials, [34][35][36][37][38] have yielded clinical candidates for Australian SMEs and international companies, and were instrumental in the discovery of antibiotics [39,40] with a novel mode of action for the biotechnology spin off company, Betabiotics.…”

Section: The Threat and Promise Of The Vastness Of Chemical Spacementioning

confidence: 99%

Adrien Albert Award: How to Mine Chemistry Space for New Drugs and Biomedical Therapies

Winkler

2015

Aust. J. Chem.

View full text Add to dashboard Cite

It is clear that the sizes of chemical, 'drug-like', and materials spaces are enormous. If scientists working in established therapeutic, and newly established regenerative medicine fields are to discover better molecules or materials, they must find better ways of probing these enormous spaces. There are essentially five ways that this can be achieved: combinatorial and high throughput synthesis and screening approaches; fragment-based methods; de novo molecular design, design of experiments, diversity libraries; supramolecular approaches; evolutionary approaches. These methods either synthesise materials and screen them more quickly, or constrain chemical spaces using biology or other types of 'fitness functions'. High throughput experimental approaches cannot explore more than a minute part of chemical space. We are nevertheless entering into an era that is data dominated. High throughput experiments, robotics, automated crystallographic beam lines, combinatorial and flow synthesis, high content screening, and the 'omics' technologies are providing a flood of data, and efficient methods for extracting meaning from it are essential. This paper describes how new developments in mathematics have provided excellent, robust computational modelling tools for exploring large chemical spaces, for extracting meaning from large datasets, for designing new bioactive agents and materials, and for making truly predictive, quantitative models of the properties of molecules and materials for use in therapeutic and regenerative medicine. We describe these broadly applicable modelling tools and provide examples of their application to serum free stem cell culture, pathogen resistant polymers for implantable devices, new markers and biological mechanisms derived from mathematical analyses of gene array data, and pharmacokinetically important physicochemical properties of small molecules. We also discuss biologically conserved peptide motifs as a design framework for small molecule drugs and give examples of the application of this concept to drug design.

show abstract

“…When we made BRNN calculation using 14 routine tests, however, the results change depending on the numbers of intermediate neurons and on switching on (or off) ARD. Rule of thumb (larger is the ratio of sample size to synaptic weights, the model works better) should be alleviated in the BRNN 9) and Roberts et al reported that whether switching on ARD function is better or not depends on the category of sample data. 7,23) Then we picked up test samples as hyperthyroid patient when they are predicted by a lot of models.…”

Section: Assisting the Diagnosis Of Thyroid Diseases With Bayesian-tymentioning

confidence: 99%

“…[1][2][3] Recently Bayesian regularized neural networks (BRNN), [4][5][6][7] which extends back-propagation learning algorithm in order to overcome its defects such as the problems of local trapping, overfitting etc. by introducing probabilistic treatment of the Bayesian inference technique for the synaptic weights, has been successfully applied for QSAR studies 8,9) even for massive sample data. …”

mentioning

confidence: 99%

Assisting the Diagnosis of Thyroid Diseases with Bayesian-Type and SOM-Type Neural Networks Making Use of Routine Test Data

Hoshi

Kawakami

Sato

et al. 2006

CHEMICAL & PHARMACEUTICAL BULLETIN

View full text Add to dashboard Cite

Both classification problems and QSAR analyses of drugs have been better studied by the multi-layer feedforward neural networks with back-propagation learning than the classical methods in pattern recognition such as linear multiple regression or adaptive least square method, since those problems often involve nonlinear relationships between descriptors and the class (/activity). [1][2][3] Recently Bayesian regularized neural networks (BRNN), [4][5][6][7] which extends back-propagation learning algorithm in order to overcome its defects such as the problems of local trapping, overfitting etc. by introducing probabilistic treatment of the Bayesian inference technique for the synaptic weights, has been successfully applied for QSAR studies 8,9) even for massive sample data. 10)Analytical chemistry is often concerned with the classification of samples into groups on the basis of chemical results while such classification process is called diagnosis if these groups are pathological cases. In the previous work 11) we analyzed the human thyroid data as the three class classification problem, that is, diagnosis by using two notable approaches, the Bayesian regularized neural networks and the self-organizing map. [12][13][14] The former presented the high classification rates and a nice soft pruning if ARD method is swichted on. The SOM enabled nicely to grasp characteric features of thyroid function from the map of well clustered thyroid patients owing to data-visualization. These results ensured that both approaches are superior to statistical analysis based on conventional multivariate analysis techniques, and would be helpful in diagnosing thyroid function based on laboratory tests. Therefore, we now try to assist the diagnosis of thyroid diseases making use of the recent clinical examination data.Conventionally, multivariate analysis has been frequently used for analyzing data on medical care.15) This kind of analysis is implicitly based on the assumption that the sample shows a normal distribution. However, this requirement is often not satisfied because of small sample size or other reasons. The relationship of the routine test data, disease history and physical findings to the diagnosis (disease name) for individual patients seems to assume a non-linear form in many cases. When neural networks are used for analysis, there is no need to assume a normal distribution of the sample. This kind of analysis is applicable even when the sample size is large or small. Furthermore, analysis using neural networks is valid for non-linear relationships.In Japan, people who go to a hospital because they are complaining of some illness usually undergo routine testing as an early step in the diagnostic process. The data from these routine tests point to one or several possible diagnoses. Full-scale diagnosis is then started with the conduct of detailed tests focused on the diseases suspected on the basis of the disease history, findings from observations and data from the routine tests. However, sometimes the findings are negati...

show abstract

Use of Automatic Relevance Determination in QSAR Studies Using Bayesian Neural Networks

Cited by 123 publications

References 40 publications

Neural Networks in 3D QSAR

Neural Networks in 3D QSAR

Adrien Albert Award: How to Mine Chemistry Space for New Drugs and Biomedical Therapies

Assisting the Diagnosis of Thyroid Diseases with Bayesian-Type and SOM-Type Neural Networks Making Use of Routine Test Data

Contact Info

Product

Resources

About