Not where you live

A starter gets treated like a zip code. The claim is that where the baker lives sorts the microbes. The authors' tables do not support that.

This is a recomputation of Landis et al., eLife 2021 (DOI 10.7554/eLife.61644), from the published source tables on Dryad (10.5061/dryad.0p2ngf1z1). It is their census, counted again. It is not a new sequencing run and not a new discovery.

The species readme and the co-occurrence effect-size identity both put the census at 500 starters. Mean annual temperature explains 0.7% of the bacterial community variation (PERMANOVA R2 = 0.007, n = 426). Residuals account for 91.7% of the sums of squares (140.677 / 153.425). The source files have no latitude or longitude, so this project draws no baker map. The paper's Mantel results stay cited: continental US r = 0.0, global yeast r = 0.07.

The co-occurrence table records Lactobacillus plantarum with L. brevis 161 times (incidences 181 and 265, expected 95.9, effect size 0.1302, Bonferroni-adjusted p = 0). That is the largest observed count among pairs in the table. The article text says 177. The poster draws 161 and says the article says 177. The gap is recorded, not averaged, and not treated as a silent correction.

Two article sentences stay cited, because the matching column is not in these workbooks: acetic acid bacteria above 1% of reads in 147 of 500 starters, and S. cerevisiae above 50% of fungal reads in 77% of samples. The species table does show S. cerevisiae in 451 starters, with a mean relative abundance of 0.761. Those are different statistics. The dough-rise correlation with total percent acetic acid bacteria (about r = -0.51 in the paper) is also cited, not redrawn. Named acetic acid bacteria in the taxon-level Spearman table are not significant on dough rise.

On the sensory subset, 40 samples have a dominant note, and 9 of those 40 are acetic acid or vinegar. Raw expert comments were dropped before any public extract. A per-pound bread price is a different series. This project does not merge with a grocery index, and it does not produce a single sourdough score.

Dashboard: https://public.tableau.com/app/profile/jason.pellerin/viz/not_where_you_live/Notwhereyoulive

Paper: https://elifesciences.org/articles/61644

Data: https://datadryad.org/dataset/doi:10.5061/dryad.0p2ngf1z1