.. where some compelling demonstration of performance/usefulness is made (ideally done in a way that is interoperable with community efforts) .. and over the longer term, communities complement/consolidate with continuous benchmarks. In principle, Omnibenchmark supports both use cases.

Interesting discussion! I agree that "extensive" benchmarks within method papers are not really the priority/essential. But at the same time, it would be tough to motivate having benchmark-less methods papers. I think we need a middle ground ..

There are some similarities, but also some differences.

Scale: cells versus regions, but probably more importantly, Feature Space: segmentation focuses a lot on DNA and cell membrane markers (possibly transcript locations), while SAC is more about multi-cellular / neighbourhoods.

Guest Editors:
Mark Robinson, University of Zurich
Fritz Joachim Sedlazeck, Baylor College of Medicine
Hong-Bin Shen, Shanghai Jiao Tong University
Jean Yee Hwa Yang, The University of Sydney
Xin Maizie Zhou, Vanderbilt University

@jeffmold.bsky.social @willmacnair.bsky.social would be very interesting indeed. Here is the classic study doing something along these lines (not in genomics):
journals.sagepub.com/doi/10.1177/...