1
00:00:00,000 --> 00:00:02,820
Hi Everyone, in this demo 
I'll introduce ChemoGraph -

2
00:00:02,820 --> 00:00:07,200
A novel visual analytics technique to 
interactively explore the chemical space.

3
00:00:09,360 --> 00:00:14,160
Here is the visual interface of the ChemoGraph. 
Users start their exploration of the chemical  

4
00:00:14,160 --> 00:00:19,140
space from the top control panel. There are 
two different ways to give inputs to the tool.

5
00:00:19,140 --> 00:00:25,680
First one is by importing files. ChemoGraph 
currently supports Json and SMI file formats.

6
00:00:25,680 --> 00:00:28,320
Importing them will set up the initial view.  

7
00:00:29,340 --> 00:00:34,260
Another way is to start by using command 
strings. We can either give an entire  

8
00:00:34,260 --> 00:00:39,120
command or simply a compound and hit 
submit, to see its scaffold class. 

9
00:00:40,320 --> 00:00:44,400
Here, I started with a lead compound 
identified from one of the high throughput  

10
00:00:44,400 --> 00:00:48,720
virtual screening studies. This string 
representation is called a SMILE string,  

11
00:00:48,720 --> 00:00:54,060
one of the common ways to represent chemical 
compounds. To get an idea of some sample compounds  

12
00:00:54,060 --> 00:01:00,000
in this scaffold class we can run an "expand" 
operation. Give a sample size and run "expand".

13
00:01:03,540 --> 00:01:10,020
Users can decompose a scaffold into sub scaffolds 
by performing an operation called "lower". Select  

14
00:01:10,020 --> 00:01:15,960
the scaffold and then the operation. Similarly, 
users can create super scaffolds by using an  

15
00:01:15,960 --> 00:01:22,320
operation called "upper". Users can perform 
these operations iteratively and can export the  

16
00:01:22,320 --> 00:01:28,080
resulting networks. This helps the users to resume 
their exploration by loading a specific state.  

17
00:01:28,920 --> 00:01:34,080
Now, I am loading a previously saved state that 
I'll be using for the reminder of this demo.

18
00:01:36,480 --> 00:01:41,520
This process of starting from A lead compound 
and elaborating it into a series of local  

19
00:01:41,520 --> 00:01:45,840
chemistries is useful in the generation of a 
Structure Activity Relationship (SAR) series,

20
00:01:45,840 --> 00:01:51,060
where current practices involve by-hand 
enumeration, which is time consuming.  

21
00:01:51,060 --> 00:01:55,560
With this visual interface users can 
do it easily while controlling the  

22
00:01:55,560 --> 00:01:58,020
series generation depending on various attributes.

23
00:01:58,620 --> 00:02:04,500
Here, we can see each scaffold is associated with 
two properties - the first one is log P value  

24
00:02:04,500 --> 00:02:11,160
and the second one is SA score. These levels 
here represent the complexity of scaffolds.

25
00:02:11,160 --> 00:02:14,340
Lower the number of rings, 
simpler the scaffolds. As  

26
00:02:14,340 --> 00:02:18,300
the level increases number of rings in 
corresponding scaffolds also increase.  

27
00:02:19,020 --> 00:02:24,600
This organization helps us in giving a spatial 
structure to the scaffolds in the chemical space.  

28
00:02:25,140 --> 00:02:30,600
Each level in ChemoGraph can be independently 
sorted by a numerical attribute that is selected  

29
00:02:30,600 --> 00:02:35,580
from the Legend. Here, log P value is 
selected. Hence, clicking on the level  

30
00:02:35,580 --> 00:02:40,080
label toggles between different ways of 
reordering, based on the log P value.

31
00:02:44,520 --> 00:02:49,380
Also, changing to SA score updates 
all the levels correspondingly.  

32
00:02:50,340 --> 00:02:54,840
Users can hover on a scaffold to 
see the directly related scaffolds.

33
00:02:57,840 --> 00:03:03,060
Selecting a scaffold highlights all the 
related scaffolds across all the levels.

34
00:03:03,060 --> 00:03:06,720
Changes in the states, either for nodes or edges,  

35
00:03:06,720 --> 00:03:09,540
can be clearly identified 
from the color highlights.

36
00:03:09,540 --> 00:03:17,280
This helps users to quickly identify both complex 
and simple scaffolds related to the selected one. 

37
00:03:18,300 --> 00:03:24,060
Zoom interaction is used to identify selected 
scaffold as a substructure in the higher levels.

38
00:03:26,580 --> 00:03:32,820
In all the upper levels of selected scaffold, 
if the selected scaffold is present,  

39
00:03:32,820 --> 00:03:38,520
zoom operation highlights the structure of the 
selected scaffold using a red color stroke. 

40
00:03:39,060 --> 00:03:42,780
One of the important aspects 
of ChemoGraph is to provide  

41
00:03:42,780 --> 00:03:48,360
the ability to take custom user trained 
biophysical property models as input,  

42
00:03:48,360 --> 00:03:55,440
and encode them on the scaffolds. By default, we 
have two common properties for each node here.  

43
00:03:56,040 --> 00:04:02,520
But users can easily add their own models through 
the API and the interface shows them here.

44
00:04:08,520 --> 00:04:14,700
So a user can select the model, and it can 
be encoded on the nodes in the network.

45
00:04:16,860 --> 00:04:22,560
Once a custom model is added onto all the 
nodes in the network, it can also be used  

46
00:04:22,560 --> 00:04:29,220
as any of the existing properties like log P or 
SA score. For example, here selecting "model 1"  

47
00:04:29,220 --> 00:04:35,880
updates all the levels and the reordering stages 
based on the selected model. Finally, once the  

48
00:04:35,880 --> 00:04:41,880
exploration is completed, users can export either 
the smile strings along with encoded properties  

49
00:04:41,880 --> 00:04:49,080
into an SMI file or export the entire view as the 
checkpoint and resume their exploration in future.

50
00:04:49,080 --> 00:04:54,000
Thank you for watching the demo of our tool 
ChemoGraph. If you want to try ChemoGraph,  

51
00:04:54,000 --> 00:04:56,280
you can access the tool at this URL.