I wish to use DESeq2 to compare the genes in two plant samples at different stages of development- A = early stage (sample A) which will contain genes for leaf, stem, shoot growth but not genes for flower growth and B = later stage (sample B) which will contain genes for leaf, stem, shoot growth and genes for flower growth. The concept is that this will enable us to identify the genes for flower growth.

BUT I have been told that DESeq2 does not work if one sample has zero flower growth genes and the other many flower growth genes-is this true?

Differential expression analysis relies on the ability to normalize for sequencing depth. A problem for all normalization methods comes when there is global changes to gene expression. A solution is either to spike in some constant RNAs or to specify which genes you know to be not changing. The latter approach can be done in DESeq2 by specifying controlGenes to the function estimateSizeFactors.