Using a reciprocal-best-hit approach, we identified 12,506 homologous introns in 11 selected species, which are designated as data set A; and we also exploited the protein ortholog annotation from Ensembl to identify 562 and 98 homologous introns in data sets that we designate B and C, respectively.
Furthermore, the significant variability observed in performance between data sets supports the use of leave-one-batch-out cross validation instead of designating a single data set as the test cohort.> -wrap-foot> Leave-one-batch-out cross validation performance for a classifier trained on three of the four whole blood infliximab data sets, and tested on the left-out data set.
This approach was preferable to designating a single data set at the test data set as it allowed us to assess the performance for each possible selection of a test data set, and allowed us to observe and average across the variability from one small test cohort to the next.
We designate this data set the EST (single map) data set, since it includes ESTs that map onto one location in the genome.
To minimize systematic errors caused by insertion/deletion events (indels), for which there is no adequate model of evolution, we also constructed a 4130 bp data set using gene coding regions only (designated as the 'Partial' data set from now on).
As noted by D'Orazio et al. (2006), there are several factors that need to be considered when designating a donor and recipient data set.
"It's a valuable data set".
The first data set, designated ER232, contained 232 chemicals, 131 active, and 101 inactive that were tested in our lab (Fang et al. 2003a).
In each cross-validation, the 80 drug-protein pairs are used as the training data while the remaining 1,324 drug-protein pairs including the 10 positive interactions are designated as the testing data set.
The data set is a 2-D data set.
This is a very rich data set.
