Example of a sentence from the dataset, annotated by 5 independent annotators (sentence 10835394_70).

Annotations in the context of the real sentence are as follows: The phenotypes of mxp19 (Fig 1B) |A2:**1SP3E3| and mxp170 (data not shown) homozygotes and hemizygotes (data not shown) are identical, |A3:**1SP3E3| |A4:**1SP3E3| |A5:**1GP3E3| suggesting that mxp19 and mxp170 are null alleles. |A1:**1SP3E3| |A2:**2SP3E1| |A3:**1SP2E0| |A4:**2SP2E0| |A5:**2GP2E3| The minimum number of sentence fragments required to represent these annotations is three: A = “The phenotypes of mxp19 (Fig 1B)” B = “and mxp170 (data not shown) homozygotes and hemizygotes (data not shown) are identical,” C = “suggesting that mxp19 and mxp170 are null alleles.” Annotators' identities are concealed with codes A1, A2, A3, A4, and A5.