The models were calibrated for each pseudo-absence dataset (the absences were weighted equally to the presences) with 10-fold cross-validation using a 70% random sample of the observed data, and model performance was assessed using the remaining 30% of the data.
The absence of dataset linkage also means that this study is unable to determine accurately the total number of ESKD cases not receiving treatment, as it is possible that some of those receiving RRT had some blood tests through Western Diagnostic Pathology even though the dominant pathology provider for these individuals would have been the public hospital based pathology service.
This was a preliminary exercise to determine the presence or absence of datasets within regions that were currently represented in SEAMAP as data poor and to identify datasets, when present, that would be considered as high priority for acquisition.
The estimated rate of false absences in the dataset (9.4%) predicts 14 false P. eos absences, 3 false P. neogaeus absences, and 4 false hybrid absences.
At 1.7-fold, it is induced more in the comparison of Hb and Tf than it is in the presence and absence of iron (Dataset S1).
Proteins harboring the interacting domains were searched for the presence or absence in this dataset and values were assigned accordingly.
Given that miR854 and miR855 are derived from retrotransposons reported to be transcriptionally reactivated in the vegetative cell [ 19], their absence from our dataset is perplexing.
We eliminated the rest 33 genomes either due to their absence in our dataset or because there was no appropriate mapping found between the KEGG ID's in HC1 and HC2 to our NCBI names in the database.
In the absence of a dataset of bona fide multifunctional proteins with which the methods could be benchmarked, we reasoned that proteins multiclustered by all three methods are the most likely candidates for multifunctionality.
Since P. eos was absent from 4 of the 45 lakes with hybrids where it was expected to occur (8.9% rate of false absences), and P. neogaeus was absent from 1 of the 10 lakes with hybrids where it was expected to occur (10% rate of false absences), the rate of false absences in the dataset was estimated to be, on average, 9.4%.
The ratio of lakes with P. eos (129 lakes) to those with P. neogaeus (23 lakes) was 5.6∶1, and this ratio ought not to have been affected by false absences in the dataset unless the rate of false absences was not equal among Phoxinus species (which is unlikely).
