Paperonium

Skin microbes · PLoS Comput Biol · 2018

The summary says exposed skin’s microbe networks are less connected. The paper’s results say more connected.

Its resultsexposed: more connectedIts chartexposed 0.176, covered 0.138Its discussionexposed: likely sturdiersame as covered skinThe summaryexposed: less connected
Its resultsexposed: more connectedIts chartexposed 0.176, covered 0.138Its discussionexposed: likely sturdiersame as covered skinThe summaryexposed: less connected

Is exposed skin’s network more or less connected than covered skin’s?

The paper says

Viruses infect bacteria. Each skin sample is drawn as a network: microbes as dots, infections as lines. The summary at the top says networks on exposed skin (open to air) are less connected than on covered skin (closed off from it), so likely more fragile.

Its own data show

Its results say the reverse: exposed skin’s networks are more connected, a gap that chance alone would give under 1 time in 1,000. Its chart agrees: on its connectedness score (higher means more connected), the typical exposed sample scores 0.176, the typical covered one 0.138. Its discussion also predicts exposed skin’s networks are sturdier.

The paper’s reasoning: the more lines tie the dots together, the less the network falls apart when a species dies out, so more connected means sturdier. Its results, chart and discussion all agree. Only the summary, the part most people read and quote, says the reverse, so readers of it alone come away thinking the wrong skin is fragile.

“The summary” means the paper’s abstract: the short summary printed at its top, which is the part most readers see.

Read the paper →Full technical write-up →

Biogeography and environmental conditions shape bacteriophage-bacteria networks across the human microbiome · PLoS Comput Biol (2018) · doi:10.1371/journal.pcbi.1006099