5 Cardio-oncology: Network-Based Prediction …
79
cardiotoxicities in the early stage of drug discovery, clinical trials, and post-marketing
surveillance.
5.2 Method and Materials
5.2.1 Computers
1. Computer requirements, laptop/desktop computers, or high-performance computing clusters with UNIX/LINUX operating systems.
2. Network analysis and visualization tools, Cytoscape (https://cytoscape.org/) or
Gephi (https://gephi.org/).
3. Python, Java, or other environments.
5.2.2 Reconstruction of Drug-Target Interaction Network
The drug-target network can be described as a bipartite graph G(D, T, P), where
the drug set denotes as D = {d 1 , d 2 , . . . , d n } , target set as T = {t 1 , t 2 , . . . , t m } , and
interaction set as P =
p i j , d i ∈ D, t j ∈ T
. An interaction is drawn between d i and
t j when drug d i binds with target t j with binding affinity (such as IC 50 , K i , or K d )
less than a given threshold value. Mathematically, a drug-target bipartite network
can be presented by an n × m adjacent matrix
p i j
, where p i j = 1 if the binding
affinity between d i and t j is less than 10 µM, otherwise p i j = 0, as described in
Eq. (5.1).
p i j =
1 IC 50 (K i ) ≤ 10 µM
0 IC 50 (K i ) > 10 µM
(5.1)
In general, we can collect drug-target interaction information from the DrugBank database (v4.3) [53], the Therapeutic Target Database (TTD) [54], and the
PharmGKB database [55]. Specifically, bioactivity data for drug-target pairs can be
collected from ChEMBL (v20) [56], BindingDB [57], and IUPHAR/BPS Guide to
PHARMACOLOGY [58]. Detailed drug-target databases are provided in Table 5.1.
To improve the quality of data, we usually focus on physical drug-target interactions based on the following three criteria: (i) The human target is represented
by a unique UniProt accession number; (ii) the target is marked as ‘reviewed’ in
the UniProt database [59]; and (iii) binding affinities, including K i , K d , IC 50, or
EC 50 each ≤10 µM. In addition, we can also build functional drug-gene association
networks from drug-induced transcriptomics data or proteomics data derived from
human cells (Table 5.1).
79
cardiotoxicities in the early stage of drug discovery, clinical trials, and post-marketing
surveillance.
5.2 Method and Materials
5.2.1 Computers
1. Computer requirements, laptop/desktop computers, or high-performance computing clusters with UNIX/LINUX operating systems.
2. Network analysis and visualization tools, Cytoscape (https://cytoscape.org/) or
Gephi (https://gephi.org/).
3. Python, Java, or other environments.
5.2.2 Reconstruction of Drug-Target Interaction Network
The drug-target network can be described as a bipartite graph G(D, T, P), where
the drug set denotes as D = {d 1 , d 2 , . . . , d n } , target set as T = {t 1 , t 2 , . . . , t m } , and
interaction set as P =
p i j , d i ∈ D, t j ∈ T
. An interaction is drawn between d i and
t j when drug d i binds with target t j with binding affinity (such as IC 50 , K i , or K d )
less than a given threshold value. Mathematically, a drug-target bipartite network
can be presented by an n × m adjacent matrix
p i j
, where p i j = 1 if the binding
affinity between d i and t j is less than 10 µM, otherwise p i j = 0, as described in
Eq. (5.1).
p i j =
1 IC 50 (K i ) ≤ 10 µM
0 IC 50 (K i ) > 10 µM
(5.1)
In general, we can collect drug-target interaction information from the DrugBank database (v4.3) [53], the Therapeutic Target Database (TTD) [54], and the
PharmGKB database [55]. Specifically, bioactivity data for drug-target pairs can be
collected from ChEMBL (v20) [56], BindingDB [57], and IUPHAR/BPS Guide to
PHARMACOLOGY [58]. Detailed drug-target databases are provided in Table 5.1.
To improve the quality of data, we usually focus on physical drug-target interactions based on the following three criteria: (i) The human target is represented
by a unique UniProt accession number; (ii) the target is marked as ‘reviewed’ in
the UniProt database [59]; and (iii) binding affinities, including K i , K d , IC 50, or
EC 50 each ≤10 µM. In addition, we can also build functional drug-gene association
networks from drug-induced transcriptomics data or proteomics data derived from
human cells (Table 5.1).