Yes, in principle there are no restrictions on the types of comparative analyses we can perform between single-cell imaging (
#singlecellimaging) and single-cell sequencing (
#singlecellsequencing) platforms. We have the flexibility to do a wide range of comparisons…from straightforward metrics like cost per gene detected, to more nuanced ones such as cost per gene not detected (i.e., missed targets with potential biological significance), or cost per gene detected but ultimately irrelevant to the study question.
We went into that rabbit hole and in the end, it’s a matter of what metrics helps you decide what way to go.
While sequencing is often marketed as “unbiased”, in practice, that’s a bit of a misnomer…the readout reflects biases in capture efficiency, amplification, and transcript abundance. Imaging, by contrast, is targeted by design, but that targeting enables higher confidence and spatial fidelity per transcript at much lower noise levels. So, comparisons must go beyond simple detection rates…we need to consider biological relevance, false discovery rates, and cost-efficiency across practical use cases.