Yes, normalization of the genus abundance matrix is typically performed during enterotyping using the Dirichlet multinomial mixture (DMM) method. One common normalization technique is to scale the genus abundances to relative proportions within each sample.
Here's how the normalization process works:
Relative Abundance Calculation: For each sample, the absolute abundance counts of each genus are divided by the total sum of abundance counts in that sample. This yields a vector of relative abundances, where each genus's abundance is expressed as a proportion of the total abundance within the sample.
Normalization Across Samples: Optionally, the relative abundance vectors across all samples can be further normalized to ensure comparability and consistency across samples. This may involve scaling the relative abundances to sum to one for each sample or applying additional normalization methods to account for differences in sequencing depth or other technical biases.
Handling Zero Abundances: Zero-inflated taxa are often encountered in microbiome data, where some genera may be absent or present at very low abundance in certain samples. Various techniques can be applied to handle zero counts, such as adding pseudocounts or employing zero-inflated models within the DMM framework.
By normalizing the genus abundance matrix, the DMM method can effectively account for differences in sequencing depth, sample composition, and technical variability, allowing for robust and accurate identification of microbial community structures and enterotypes.
Please follow me if it's helpful. All the very best. Regards, Safiul
Thank you very much for your reply. If I understand correctly, normalization is not included in the DMN function. We need to perform normalization on the database beforehand. The DMN function requires a database with counts of OTU.