Use of biclustering for missing value imputation in gene expression data

Cheng, Kin-On; Law, Ngai-Fong; Siu, Wan-Chi

doi:10.5430/air.v2n2p96

Cited by 14 publications

(12 citation statements)

References 30 publications

(43 reference statements)

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…This indicates that the number of station-clusters and of timestamps-clusters should be optimized in future studies. For example, mean squared residue (MSR) (Cheng andChurch 2000, Zhou andAshfaq 2006) is an index that can be used to evaluate clustering results and thereby optimize cluster numbers. (2) The current BBAC_I algorithm can only deal with a single numeric value for each element in the data matrix and consequently analyses only one observed propertyyearly, monthly and daily averaged temperatures in this study.…”

Section: Discussionmentioning

confidence: 99%

Co-clustering geo-referenced time series: exploring spatio-temporal patterns in Dutch temperature data

Zurita-Milla

Kraak

2015

International Journal of Geographical Information Science

View full text Add to dashboard Cite

Clustering allows considering groups of similar data elements at a higher level of abstraction. This facilitates the extraction of patterns and useful information from large amounts of spatio-temporal data. Till now, most studies have focused on the extraction of patterns from a spatial or a temporal aspect. Here we use the Bregman block average co-clustering algorithm with I-divergence (BBAC_I) to enable the simultaneous analysis of spatial and temporal patterns in geo-referenced time series (time evolving values of a property observed at fixed geographical locations). In addition, we present three geovisualization techniques to fully explore the co-clustering results: heatmaps offer a straightforward overview of the results; small multiples display the spatial and temporal patterns in geographic maps; ringmaps illustrate the temporal patterns associated to cyclic timestamps. To illustrate this study, we used Dutch daily average temperature data collected at 28 weather stations from 1992 to 2011. The co-clustering algorithm was applied hierarchically to understand the spatio-temporal patterns found in the data at the yearly, monthly and daily resolutions. Results pointed out that there is a transition in temperature patterns from northeast to southwest and from 'cold' to 'hot' years/months/days with only 3 years belonging to 'cool' or 'cold' years. Because of its characteristics, this newly introduced algorithm can concurrently analyse spatial and temporal patterns by identifying location-timestamp co-clusters that contain values that are similar along both the spatial and the temporal dimensions.

show abstract

Section: Discussionmentioning

confidence: 99%

Co-clustering geo-referenced time series: exploring spatio-temporal patterns in Dutch temperature data

Zurita-Milla

Kraak

2015

International Journal of Geographical Information Science

View full text Add to dashboard Cite

show abstract

“…Cheng et al [21] use linear interpolation in gene expression analysis, which is a specialized area of research that cannot be placed into the categories of regression or classification. It is characterised by scarcity of data, and no clear distinction exists between instances and attributes.…”

Section: Related Workmentioning

confidence: 99%

Diagnostic with incomplete nominal/discrete data

Jelinek¹,

Yatsko²,

Stranieri³

et al. 2015

AIR

View full text Add to dashboard Cite

Missing values may be present in data without undermining its use for diagnostic / classification purposes but compromise application of readily available software. Surrogate entries can remedy the situation, although the outcome is generally unknown. Discretization of continuous attributes renders all data nominal and is helpful in dealing with missing values; particularly, no special handling is required for different attribute types. A number of classifiers exist or can be reformulated for this representation. Some classifiers can be reinvented as data completion methods. In this work the Decision Tree, Nearest Neighbour, and Naive Bayesian methods are demonstrated to have the required aptness. An approach is implemented whereby the entered missing values are not necessarily a close match of the true data; however, they intend to cause the least hindrance for classification. The proposed techniques find their application particularly in medical diagnostics. Where clinical data represents a number of related conditions, taking Cartesian product of class values of the underlying sub-problems allows narrowing down of the selection of missing value substitutes. Real-world data examples, some publically available, are enlisted for testing. The proposed and benchmark methods are compared by classifying the data before and after missing value imputation, indicating a significant improvement.

show abstract

“…Eight different biclustering methods were applied to the combined datasets, namely Cheng and Church (CC) [ 15 ], Plaid [ 16 ], Bimax [ 17 ], Spectral [ 33 ], FLOC [ 34 ], XMOTIFS [ 35 ], large average sub-matrices (LAS) [ 36 ], bipartite spectral graph partitioning (BSGP) [ 37 ]. At all genome sizes, Spectral and XMOTIFS produced no clusters, while CC produced a single trivial cluster that encompasses the entire genome and all of the data samples.…”

Section: Resultsmentioning

confidence: 99%

“…Although, to the authors’ knowledge, this research question has not been answered in an unsupervised way previously, it has been raised and discussed implicitly and explicitly in many studies [ 2 – 4 , 12 – 14 ]. However, biclustering methods, such as Cheng and Church (CC) [ 15 ], Plaid [ 16 ], Bimax [ 17 ], and others, mine a data matrix for the rows (corresponding to genes) that show consistent co-expression across all or some of the matrix columns (corresponding to samples). Although such methods were designed to mine a single dataset, multiple datasets may be concatenated to provide a single data matrix that is fed to biclustering analysis.…”

Section: Introductionmentioning

confidence: 99%

UNCLES: method for the identification of genes differentially consistently co-expressed in a specific subset of datasets

Abu‐Jamous

Roberts

et al. 2015

BMC Bioinformatics

View full text Add to dashboard Cite

BackgroundCollective analysis of the increasingly emerging gene expression datasets are required. The recently proposed binarisation of consensus partition matrices (Bi-CoPaM) method can combine clustering results from multiple datasets to identify the subsets of genes which are consistently co-expressed in all of the provided datasets in a tuneable manner. However, results validation and parameter setting are issues that complicate the design of such methods. Moreover, although it is a common practice to test methods by application to synthetic datasets, the mathematical models used to synthesise such datasets are usually based on approximations which may not always be sufficiently representative of real datasets.ResultsHere, we propose an unsupervised method for the unification of clustering results from multiple datasets using external specifications (UNCLES). This method has the ability to identify the subsets of genes consistently co-expressed in a subset of datasets while being poorly co-expressed in another subset of datasets, and to identify the subsets of genes consistently co-expressed in all given datasets. We also propose the M-N scatter plots validation technique and adopt it to set the parameters of UNCLES, such as the number of clusters, automatically. Additionally, we propose an approach for the synthesis of gene expression datasets using real data profiles in a way which combines the ground-truth-knowledge of synthetic data and the realistic expression values of real data, and therefore overcomes the problem of faithfulness of synthetic expression data modelling. By application to those datasets, we validate UNCLES while comparing it with other conventional clustering methods, and of particular relevance, biclustering methods. We further validate UNCLES by application to a set of 14 real genome-wide yeast datasets as it produces focused clusters that conform well to known biological facts. Furthermore, in-silico-based hypotheses regarding the function of a few previously unknown genes in those focused clusters are drawn.ConclusionsThe UNCLES method, the M-N scatter plots technique, and the expression data synthesis approach will have wide application for the comprehensive analysis of genomic and other sources of multiple complex biological datasets. Moreover, the derived in-silico-based biological hypotheses represent subjects for future functional studies.Electronic supplementary materialThe online version of this article (doi:10.1186/s12859-015-0614-0) contains supplementary material, which is available to authorized users.

show abstract

Use of biclustering for missing value imputation in gene expression data

Cited by 14 publications

References 30 publications

Co-clustering geo-referenced time series: exploring spatio-temporal patterns in Dutch temperature data

Co-clustering geo-referenced time series: exploring spatio-temporal patterns in Dutch temperature data

Diagnostic with incomplete nominal/discrete data

UNCLES: method for the identification of genes differentially consistently co-expressed in a specific subset of datasets

Contact Info

Product

Resources

About