摘要
Objective The quality of microarray data influences the accuracy of comparative genomic analyses to a large extent.To ensure that the results obtained by using an in situ synthesized microarray are accurate,data quality is to be assessed by evaluating the melting temperature (Tm) of probes,probability of false synthesis rates,and fragmentation of labeled targets.Methods DNA from the Yersinia pestis vaccine strain EV76 was used for microarray analyses.Microarray results were confirmed by PCR.Statistical and bioinformatics methods were employed to perform microarray data analyses and evaluation.Results Correlation coefficients of the three datasets were above 0.95 after two-time stripping and hybridization with a labeled DNA with the size of fragmentation being 200 bp-2 kb,which showed that the hybridization results were highly reproducible.Correlation coefficients were lower with the values ranging from 0.87 to 0.92 between the datasets generated from hybridization with different sizes of the labeled DNA fragment.For the relationship between Tm and signal intensity,there was a different distribution of Tm in the lowest 300 or 3 000 probes with a range of 70 ℃-72 ℃ and the highest 300 or 3 000 probes with a range of 72 ℃-74 ℃.Conclusion The results of this study suggest that the initial microarray design may affect the accuracy of final analyses and that the probe Tm and the size of the labeled fragment may be the two factors of the greatest importance.
Objective The quality of microarray data influences the accuracy of comparative genomic analyses to a large extent.To ensure that the results obtained by using an in situ synthesized microarray are accurate,data quality is to be assessed by evaluating the melting temperature (Tm) of probes,probability of false synthesis rates,and fragmentation of labeled targets.Methods DNA from the Yersinia pestis vaccine strain EV76 was used for microarray analyses.Microarray results were confirmed by PCR.Statistical and bioinformatics methods were employed to perform microarray data analyses and evaluation.Results Correlation coefficients of the three datasets were above 0.95 after two-time stripping and hybridization with a labeled DNA with the size of fragmentation being 200 bp-2 kb,which showed that the hybridization results were highly reproducible.Correlation coefficients were lower with the values ranging from 0.87 to 0.92 between the datasets generated from hybridization with different sizes of the labeled DNA fragment.For the relationship between Tm and signal intensity,there was a different distribution of Tm in the lowest 300 or 3 000 probes with a range of 70 ℃-72 ℃ and the highest 300 or 3 000 probes with a range of 72 ℃-74 ℃.Conclusion The results of this study suggest that the initial microarray design may affect the accuracy of final analyses and that the probe Tm and the size of the labeled fragment may be the two factors of the greatest importance.
基金
supported by a grant from the National High Technology Research and Development Program of China(863 Program,No. 2006AA2Z4A7)
