Hello all,
I was wondering about the citation when using diversity core-metrics to calculate Bray-Curtis dissimilarity. The citation given by QIIME 2 is Sørensen (1948), but wouldn't Bray and Curtis (1957), DOI: 10.2307/1942268, be more appropriate?
2 Likes
Lol, yeah I guess it makes sense to cite the paper written by Bray and Curtis.
Sørensen is older and funky: unlike other popular beta diversity numbers (Unifrac), it's a 'similarity measure' and not a 'distance metric,' so a high Sørensen–Dice index means samples are similar. It's also binary, not weighted by feature abundances.
But what if you wanted Sørensen to be weighted by abundance and be a dissimilarity like the other numbers?
Spoiler
Wikipedia: Sørensen–Dice coefficient, Abundance_version
- Bray–Curtis similarity (1 minus the Bray-Curtis dissimilarity )
I'm not sure what's best to cite.
4 Likes