Paperonium

Child liver disease · PLoS Genet · 2013

“With one DNA type, child liver disease is 29% less likely.” Its counts: 55% to 100%.

The paper’s table and textIts own counts:0 of 71 sick children29%23%34%55 to 100%0%20%40%60%80%100%no difference
The paper’s tableand textIts own counts:0 of 71 sickchildren29%23 to 34%55 to 100%0%50%100%no difference

Percent less likely to have the disease, for people with DNA type B4

The paper says

B4 is one version of a small piece of DNA passed only from mother to child. The paper’s main genetic finding: people with B4 are 29% less likely to have biliary atresia, a liver disease of young children, with a range of 23% to 34%.

Its own data show

Its own table finds B4 in 23 of 200 healthy children and adults, about 1 in 9, but in none of 71 sick children (1 in 9 would be about 8). From those counts, the usual small-number calculation gives 55% to 100% less likely: with no sick child carrying B4, 100% can’t be ruled out.

A range is the set of values the data can’t rule out; few cases make it wide. The link itself holds up, but the paper’s 29% measures something else: 71% of people without B4 were in the healthy group, against 100% of those with it (71 is 29% below 100). Anyone reusing it gets the wrong size and false certainty.

“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 →

Associations of Mitochondrial Haplogroups B4 and E with Biliary Atresia and Differential Susceptibility to Hydrophobic Bile Acid · PLoS Genet (2013) · doi:10.1371/journal.pgen.1003696