Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

An algorithm for automatic evaluation of the spot quality in two-color DNA microarray experiments.

Eugene Novikov1, Emmanuel Barillot

  • 1Service Bioinformatique, Institut Curie, 26 Rue d'Ulm, 75248 Paris Cedex 05, France. Eugene.Novikov@curie.fr

BMC Bioinformatics
|December 13, 2005
PubMed
Summary
This summary is machine-generated.

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Mesenchymal to epithelial transition (MET) in cancer progression: insights from logical modeling.

Scientific reports·2026
Same author

Multiomic integration reveals tumoral heterogeneity of lipid dependence within lethal group 3 medulloblastoma.

Cancer cell·2026
Same author

sCellST predicts single-cell gene expression from H& E images.

Nature communications·2026
Same author

NeKo: A tool for automatic network construction from prior knowledge.

PLoS computational biology·2025
Same author

Author Correction: Systems medicine disease maps: community-driven comprehensive representation of disease mechanisms.

NPJ systems biology and applications·2025
Same author

Cell-type deconvolution methods for spatial transcriptomics.

Nature reviews. Genetics·2025

This article introduces a new automated computational method to assess the reliability of individual spots on DNA microarray images, helping researchers identify and filter out low-quality data points to improve the accuracy of gene expression analysis.

Area of Science:

  • Genomics and bioinformatics research within DNA microarray technology
  • Computational biology and statistical modeling of gene expression data

Background:

No prior work had resolved the persistent challenge of accurately assessing the reliability of individual data points within high-throughput genomic imaging. DNA microarray technologies allow for the simultaneous quantitative characterization of thousands of genes. However, the quality of obtained experimental data is often far from ideal. This gap motivated the development of robust computational frameworks to ensure data integrity. Prior research has shown that measured images represent a regular collection of spots. The intensity of light at each spot corresponds to the DNA copy number or gene expression level. That uncertainty drove the need for automated quality control at the level of individual spot identification. The problem remains difficult to formalize due to the diversity of instrumental and biological factors influencing results.

Purpose Of The Study:

The aim of this study is to develop an algorithm for the automatic evaluation of spot quality in two-color DNA microarray experiments. Researchers seek to address the difficulty of formalizing quality control due to diverse instrumental factors. This work focuses on creating a model that combines multiple quality characteristics into a single value. The authors intend to provide a tool that quantifies the reliability of individual spots. They aim to improve the accuracy of gene expression measurements by identifying deficient spots. The study addresses the need for systematic assessment at the level of individual identification. By doing so, the researchers hope to enhance the interpretation of differential gene expression. This effort is motivated by the requirement to eliminate or weight unreliable data in subsequent analysis.

Keywords:
gene expressionfluorescence intensitycomputational biologydata quality control

Frequently Asked Questions

The researchers propose a model that calculates an overall quality value by combining several defined characteristics. This score helps distinguish reliable signals from poor ones, allowing for the exclusion or weighting of specific data points based on their calculated integrity.

The authors utilize a training procedure that leverages information from replicated spots. By comparing these replicates, the tool identifies discrepancies, as unspoiled spots should exhibit very close intensity ratios, while defective ones show greater variability.

The researchers state that quality control must occur at the level of individual spot identification. This granularity is necessary because instrumental and biological factors vary significantly across the array, making global assessment insufficient for accurate gene expression quantification.

Related Experiment Videos

Main Methods:

The review approach involves developing a training procedure to evaluate the contribution of individual characteristics to overall quality. Researchers define a set of quality metrics to assess each spot. They utilize information available from replicated spots to calibrate the model. These replicates exist either within the same array or across a set of replicated arrays. The design assumes that unspoiled spots must yield very close ratios. Conversely, poor spots produce greater diversity in ratio estimates. This computational framework provides an automatic tool for quantifying signal reliability. The approach focuses on identifying various types of deficiency occurring in the technology.

Main Results:

Key findings from the literature indicate that the developed procedure successfully quantifies spot quality automatically. The model estimates the ratio of measured fluorescence intensities to reveal differential gene expression. It effectively identifies different types of spot deficiency that commonly occur. The results show that unspoiled replicated spots exhibit very close ratio estimates. In contrast, poor spots yield significantly greater diversity in these estimates. The researchers established a training procedure to evaluate the contribution of each individual characteristic. Quality values assigned to each spot facilitate the elimination of unreliable data. Finally, these values allow for the weighting of ratio estimates in follow-up analysis procedures.

Conclusions:

The authors propose an automated tool to quantify spot quality and identify various types of deficiency. This procedure provides a systematic way to assess reliability in microarray experiments. Quality values assigned to each spot allow researchers to eliminate unreliable data points. Alternatively, these values can weight the contribution of each ratio estimate in subsequent analysis. The researchers suggest that this approach improves the overall accuracy of gene expression studies. By utilizing replicated spots, the model effectively distinguishes between high-quality and poor-quality signals. The study demonstrates that consistent ratios across replicates indicate reliable measurements. This synthesis implies that automated quality control is a viable strategy for enhancing genomic data interpretation.

The authors use replicated spots located within the same array or across multiple arrays to provide the necessary data for training. This replication serves as the ground truth to determine which spots are reliable versus those that are deficient.

The study measures the ratio of fluorescence intensities between test and control samples. Poor spots are identified by their tendency to yield greater diversity in these ratio estimates compared to the consistent values observed in high-quality replicates.

The authors claim that their automated procedure allows for the identification of different types of spot deficiency. This capability enables researchers to refine follow-up analysis by either removing flawed data or adjusting the influence of specific ratio estimates.