Biology Reference
In-Depth Information
Number of proteins
Success rate (%)
120
100
100
94
79
77
80
75
72
67
60
60
56
43
40
33
27
20
19
20
16
6
5
5
1
0
0
4
5
6
7
8
9
10
11
12
14
Number of clusters
Fig. 2.6
The metaPocket prediction success rates on the drug-target data-set at the top 3 versus the
number of clusters (meta-pocket sites). The number of proteins is also indicated
approaches. In general, one pocket site has higher probability to be a real ligand
binding site when it is picked out by multiple methods as top predictions. This is not
surprising because different pocket detection methods use different scoring func-
tions to rank these cavities and metaPocket clusters all the identified pocket sites
according to their spatial distance and re-ranks them by summing up the z-scores of
different methods.
In the combining procedure of metaPocket, only the top 3 pocket sites from each
of 8 single methods are taken into account and these 24 pocket sites are clustered
into different clusters (so called meta-pocket site) according to their spatial similarity.
In the evaluation of metaPocket on the drug-target dataset, the number of fi nal clusters
for each protein and the prediction success rates of metaPocket on those proteins
are quite diverse. Figure
2.6
shows the distribution of the number of proteins with
different number of clusters on the drug-target dataset, and the success rates for
those proteins having the same number of clusters. Overall, the number of clusters
ranges from 4 to 14, which means there are 4-14 cavities (meta-pocket sites) on
protein surfaces generally. It is note that there are 5 cases, in which those 24 pockets
are clustered into 4 clusters, meaning that those 5 proteins only have 4 big cavities
on their surfaces and all the 8 methods correctly picked them up at their top 3 predic-
tions. In these five cases, metaPocket all predicted the ligand-binding sites correctly.
In contrast, there is only one case that the number of final clusters is 14, which indicates
that this protein has 14 cavities on its surface and each of 8 methods picked up
different pockets at their top 3 predictions. The real ligand binds to one of those 14
cavities and metaPocket failed to recognize it correctly at the top 3 predictions.
As shown in Fig.
2.6
, most of the proteins have 7 (43 cases) or 8 (56 cases) cavities
Search WWH ::
Custom Search