Automatic Teaching–Learning-Based Optimization: A Novel Clustering Method for Gene Functional Enrichments - Computational Intelligence Techniques for Comparative Genomics - page 30

5.2 Experiment 2

This experiment focuses mainly to attain the gene functional enrichment of the
incurred clusters of yeast datasets in Table 1 and re-validate the outcomes of
AutoTLBO in experiment 1. The outputs of experiment 1 are used as inputs to the
experiment 2. The Web-based functional annotation tools such as FatiGo [ 23 ] and
DAVID [ 24 ] are used, particularly for gene-enrichment analysis. The clustered gene-
IDs of each yeast datasets is segmented individually into two lists of genes, i.e., a
group of interest as foreground and rest of genes as background lists. These lists are
passed as inputs to the aforementioned tools for gene functional enrichment. The GO
biological process is triggered on the gene lists to obtain signi cant results in range

of selected level of gene ontology, gene annotations, p-value, etc. Since experiment
1 uses the yeast datasets, the genomics organism is treated as S. cerevisiae.

5.2.1 Outputs of FatiGo

Table 2 is inferred with the GO biological process applied between the levels 3 9

-

on all the ve comparing datasets. The percentages of annotations produced by

FatiGo in list 1 and list 2 of yeast234 and yeast384 datasets are outstanding. In the
list 1 of yeast3 and yeast5, the percentage of annotations are 72 and 78.23 %,
respectively, and in the list 2 of yeast3 and yeast5, the percentage of annotations are
73.48 and 77.66 %, respectively. This indicates that FatiGo has marked a reason-
able good percentage of annotations on the list. The reason that was most looming
was of the enormous size of data. The experimental results of Yeast2946 are not
shown in this paper since the obtained results are merely equivalent to yeast2885.
The clustering accuracy of TLBO was once again proven from column 6 and 7 of
Table 2 . The percentage of misclassi

cation of the algorithm is zero in both in

Table 2 GO biological process on yeast datasets
GO biological process (levels from 3 to 9)
Dataset Total

genes

ID annotations Duplicate management Number of

signi cant
terms

List 1
annotations

List 2
annotations

List 1
duplicates

List 2
duplicates

Yeast1 238 109 of 113

(96.4 %)

101 of 104
(97.12 %)

16 of 129
(0.12 %)

4 of 108
(0.04 %)

75

Yeast2 384 10 of 131

(83.97 %)

155 of 190
(81.58 %)

0 of 131
(0 %)

0 of 190
(0 %)

11

Yeast3 2,885 956 of

1,327
(72 %)

1,144 of
1558
(73.48 %)

0of
1,327
(0 %)

0of
1,558
(0 %)

6

Yeast5 4,382 1,193 of

1,525
(78.23 %)

2,218 of
2,856
(77.66 %)

0of
1,525
(0 %)

0of
2,856
(0 %)

4