This project explores cancer treatment evolution through an analysis of cancer-related data from BindingDB, focusing on how research trends have changed over time and what these trends reveal about treatment progress. This project aims at understanding relationships between ligands and proteins to extract valuable information on cancer treatment. By investigating how the proportions of treatments vary among cancer types, as well as the impact of treatments targeting mutant proteins, we assess how current approaches address challenges posed by cancer mutations, pathways, and side effects. Our goal is to tell the story of cancer treatment advancements, identifying pivotal research milestones and revealing insights on the struggles due to side effects. This work ultimately illustrates the role of molecular interaction data in advancing cancer research and addressing critical health challenges.
- How has the number of treatments evolved through time?
- Are there any cancer-related proteins that rapidly became major research targets?
- Which cancer-related proteins are most frequently targeted by therapeutic drugs, based on BindingDB and DrugBank data?
- How many mutants does every cancer-related protein have, and how effective are the ligands against these mutants?
- What are the key ligands that ultimately get approved as drugs, and what are there binding/inhibiting/concatring/kinetic properties?
- How do mutations in cancer-related proteins influence the binding affinity of therapeutic drugs and treatment success, knowing that cancer cells mutate to gain treatment resistance?
- What are the side effects of ligands that show great properties against cancer ?
While BindingDB will serve as the primary dataset, DrugBank provides valuable information on treatment approval, side effects, and other pharmacological properties of the ligands.
- Source: DrugBank (https://www.drugbank.ca/)
- Size and Format: Structured data (XML) with detailed drug information, including targets, mechanisms of action, and clinical uses.
- Description: DrugBank contains extensive data on over 13,000 drugs and their interactions with biological targets. Integrating DrugBank data with BindingDB allows us to explore how specific drugs interact with cancer-related proteins, aiding in the identification of potential therapeutic compounds. Additionally, Drugbank contains the information on approved drugs and hence, ligands that were deemed potent enough, as well their side effects.
- Plan for Management: Data from DrugBank will be mapped to relevant proteins from BindingDB based on their interaction data (unique names, or structure), allowing us to identify high-affinity binding drugs and evaluate their effectiveness in treating cancer. Additional sources for context and story-telling : SEER, COSMIC, TCGA, IARC
- Clean and preprocess data from BindingDB, DrugBank,
- Map Protein IDs in BindingDB to those of DrugBank to integrate drug-target interactions contained in BindingDB to clinical data from DrugBank. IDs used :
- PubChem CID: PubChem, CID, Compound,
- ChEBI ID of Ligand: ChEBI, Ligand, Bioentity,
- ChEMBL ID of Ligand: ChEMBL, Ligand, Bioactivity,
- DrugBank ID of Ligand: DrugBank, Drug, Pharmaceutical,
- KEGG ID of Ligand: KEGG, Ligand, Pathway,
- ZINC ID of Ligand: ZINC, Ligand, Screening,
- Ligand SMILES: SMILES, Structure, Notation
- Ligand InChI Key: InChI Key, Structure, Identifier
- BindingDB MonomerID: BindingDB, Monomer, Interaction,
- Merge BindingDB, DrugBank to and save the newly formed dataset : Merged_DB .
- Filter the Merged_DB based on its “specific-function”, “Target Name”, “entry name of target chain” attributes (for proteins) by identifying cancer-related keywords.
- Look into correlations between all attributes of interest, in particular binding properties, temperature and pH.
- Visualize the distribution of binding affinities for cancer-related proteins and the drugs associated with them using histograms.
- Perform trend analysis by looking at the proteins present in published articles using their DOI.
- Link proteins to their mutants based on their target names and find all ligands that were tested on these proteins and have properties above a certain threshold.
- Compare chemical properties of ligands, and identify which ones are present in approved medical drugs.
- Analyze their side effects, by looking at the other molecules they bind to.
- Create interactive visualizations using tools like Plotly, Seaborn, or Matplotlib to display protein-ligand interactions, treatment efficacy, and survival outcomes.
- Use network diagrams to represent protein-drug interactions (4).
Week 10 : Identify which hypotheses are true, while analyzing time trends, and proteins (visualization) Week 11 : Network Analysis, Ligand Analysis (visualization) Week 12 : Side effects (visualization) Week 13 : Conclude and present the storyline on our website.
Each of us is responsible for one topic. Each week, we talk about our findings and present our visualizations.
Find data on cancer cases, deadliness over the years, and trends in general Get access to DrugBank and understand how to link it to BindingDB Create template to read BindingDB efficiently > mySQL > tsv Preprocessing of the data - data cleaning Identify proteins, which are related to cancer Get a better understanding on Chemical cancer treatment to understand the main metabolic pathways Define all cancer-related parameters that could show interesting trends over the years and will help us lay out the story of the fight against cancer