Normalization and centering of array-based heterologous genome hybridization based on divergent control probes

Area of Science:

• Genomics
• Bioinformatics
• Molecular Biology

Background:

• Heterologous nucleic acid hybridization on model-organism arrays is a proposed genomic resource for non-model organisms.
• Conventional normalization methods (e.g., quantile normalization) are unsuitable for non-specific hybridization due to distribution assumptions.
• A novel algorithm is proposed to normalize and center intensity data from heterologous hybridization without prior distribution assumptions.

Purpose of the Study:

• To develop and validate a new algorithm for normalizing and centering genomic data from heterologous hybridization in non-model organisms.
• To reduce false positive probe matches and assess the suitability of heterologous hybridization for genomic analysis.
• To improve the estimation of sequence variation and gene expression in non-model organisms.

Main Methods:

• Normalization adjusts for Gibbs free energy binding.
• Centering adjusts for the median of control probes assumed to be equivalently dissimilar across species.
• The algorithm was compared against Loess and quantile normalization using data from two related nematode species.

Main Results:

• The proposed algorithm successfully detected sequence variations (deletions).
• It outperformed quantile normalization in reducing false positive probe matches between Caenorhabditis elegans and C. briggsae.
• Despite improvements, probe fluorescence intensity showed poor correlation with sequence similarity for reliable homology inference.

Conclusions:

• Cross-species hybridization is a viable method for adapting genome-enabled tools for closely related non-model organisms.
• Appropriate normalization and centering are crucial for handling nucleic acid hybridization with diverged sequences.
• For short probes (25-mer), hybridization intensity alone may not reliably indicate sequence similarity for homology inference.