Article (Scientific journals)
In search of functional association from time-series microarray data based on the change trend and level of gene expression.
He, Feng; Zeng, An-Ping
2006In BMC Bioinformatics, 7, p. 69
Peer Reviewed verified by ORBi
 

Files


Full Text
In search of functional association from time-series microarray data 2006.pdf
Publisher postprint (984.39 kB)
Download

The original publication is available at http://www.biomedcentral.com/1471-2105/7/69


All documents in ORBilu are protected by a user license.

Send to



Details



Keywords :
Algorithms; Cluster Analysis; Gene Expression Profiling/methods; Oligonucleotide Array Sequence Analysis/methods; Pattern Recognition, Automated/methods; Protein Interaction Mapping/methods; Time Factors
Abstract :
[en] BACKGROUND: The increasing availability of time-series expression data opens up new possibilities to study functional linkages of genes. Present methods used to infer functional linkages between genes from expression data are mainly based on a point-to-point comparison. Change trends between consecutive time points in time-series data have been so far not well explored. RESULTS: In this work we present a new method based on extracting main features of the change trend and level of gene expression between consecutive time points. The method, termed as trend correlation (TC), includes two major steps: 1, calculating a maximal local alignment of change trend score by dynamic programming and a change trend correlation coefficient between the maximal matched change levels of each gene pair; 2, inferring relationships of gene pairs based on two statistical extraction procedures. The new method considers time shifts and inverted relationships in a similar way as the local clustering (LC) method but the latter is merely based on a point-to-point comparison. The TC method is demonstrated with data from yeast cell cycle and compared with the LC method and the widely used Pearson correlation coefficient (PCC) based clustering method. The biological significance of the gene pairs is examined with several large-scale yeast databases. Although the TC method predicts an overall lower number of gene pairs than the other two methods at a same p-value threshold, the additional number of gene pairs inferred by the TC method is considerable: e.g. 20.5% compared with the LC method and 49.6% with the PCC method for a p-value threshold of 2.7E-3. Moreover, the percentage of the inferred gene pairs consistent with databases by our method is generally higher than the LC method and similar to the PCC method. A significant number of the gene pairs only inferred by the TC method are process-identity or function-similarity pairs or have well-documented biological interactions, including 443 known protein interactions and some known cell cycle related regulatory interactions. It should be emphasized that the overlapping of gene pairs detected by the three methods is normally not very high, indicating a necessity of combining the different methods in search of functional association of genes from time-series data. For a p-value threshold of 1E-5 the percentage of process-identity and function-similarity gene pairs among the shared part of the three methods reaches 60.2% and 55.6% respectively, building a good basis for further experimental and functional study. Furthermore, the combined use of methods is important to infer more complete regulatory circuits and network as exemplified in this study. CONCLUSION: The TC method can significantly augment the current major methods to infer functional linkages and biological network and is well suitable for exploring temporal relationships of gene expression in time-series data.
Disciplines :
Biochemistry, biophysics & molecular biology
Author, co-author :
He, Feng ;  Helmholtz Centre for Infection Research
Zeng, An-Ping
Language :
English
Title :
In search of functional association from time-series microarray data based on the change trend and level of gene expression.
Publication date :
2006
Journal title :
BMC Bioinformatics
ISSN :
1471-2105
Publisher :
BioMed Central, United Kingdom
Volume :
7
Pages :
69
Peer reviewed :
Peer Reviewed verified by ORBi
Available on ORBilu :
since 22 April 2013

Statistics


Number of views
51 (5 by Unilu)
Number of downloads
45 (1 by Unilu)

Scopus citations®
 
20
Scopus citations®
without self-citations
15
OpenCitations
 
17
WoS citations
 
19

Bibliography


Similar publications



Contact ORBilu