Read depth and biological replicates are different concepts.

1. Read depth means sequencing the same sample with more reads, which improves the precision of transcript quantification, especially for low-abundance genes.

2. Sample size means having more independent biological samples. This does not give you more reads from the same sample; instead, it lets you measure biological variation across samples. If the same conclusion is observed repeatedly across independent samples, it is less likely to be due to random error.

3. In most differential expression studies, B is usually more powerful than A. A: 3 samples at depth X each provides more precise measurement for each individual sample. B: 9 biological replicates at depth X/3 each provides a much better estimate of biological variability, which is usually the key factor for statistical testing.

So, once sequencing depth is already sufficient, adding biological replicates usually improves statistical power more than adding extra depth. Very low depth can still be a problem, but beyond a reasonable threshold, more biological replicates are usually more valuable than more reads per sample.