Interning at NUS

Recently, I had the privilege of interning with the N.1 institute at the National University of Singapore (NUS). They’re developing a platform called CURATE.AI, which uses artificial intelligence to recommend drug dosages for patients. My internship was led by the wonderful Dr. Agata Blasiak, Dr. Theodore Kee, and Prof. Dean Ho.

This was one of the most insightful experiences I’ve had throughout my high school career. I want to share what I’ve learned, so here is an excerpt of my notes:

Singapore’s Ageing Population & Sector Implications (Dr. Daniel Teh)

  • Problem: Singaporean population growing older & less healthy due to disease. 2 strategies:

    • Invest into researching each disease

    • Invest into ageing, as that's the fundamental factor that underlies all of these diseases

  • Increasing demand in market for ageing-related treatments, and not just in Singapore

  • A longer lifespan does not entail a longer health-span. See this article

  • How do researchers actually study ageing?

    • Use animals that have shorter lifespans than humans but are similar to them, or else it would take too long (yeast and c. elegans and mice)

    • Key recent finding: deleting the mTOR1 pathway increases yeast lifespan—can translate to humans

    • Human cells organoid models, induced pluripotent stem cells (iPSCs), any cell in the body is converted into stem cells

    • How do we measure ageing? One way is with C-reactive protein (CRP)

    • Having a clear metric shows results to investors, attracts greater funding, and creates direct connection to pharmaceutical sector

  • Rapamycin is a drug with anti-ageing effects, but also has many side effects

  • Dosage recommendation programs like CURATE.AI can potentially revolutionise pharmaceutical market, increasing the demand for drugs historically avoided due to their side effects

clinical trials and human research (Dr. Theo Kee)

  • History of human research:

    • Nuremberg in 1947

    • Helsinki Declaration in 1964

    • National Research Act of 1974: US department sets forth regulations protecting human subjects; comes after Tuskegee experiment (which exposed many complex issues w/ human research)

    • Belmont's principles in 1979

  • Animal and human research ethics are a big problem

    • Ethics of human research go back to ideas of inherent superiority, like racial superiority and eugenics

    • Institution Review Board (IRB) manages these problems

    • IRB reviews clinical trials, when treatments (drug or otherwise) are tested on a population (humans or otherwise). Important things in clinical trials:

      • Safety of the participant

      • Informed consent

      • Confidentiality

  • What is a clinical trial and what are its stages? What is the process of drug approval after a clinical trial?

Python Data Analysis

  • Data analysis is an important skill because a lot of data and big data—information overflow

  • If we have a data file (like an Excel sheet), we can do all sorts of things with it

    • We can import the data using Pandas, and define a variable as this imported data (that variable is called a dataframe)

    • We can plot the data w/ help of Matplotlib, Pandas, or Seaborn

    • We can add flair to our document using Markdown

  • I made a cheat sheet of commands to know when manipulating datasets: 

# DEFINE A VARIABLE AS A DATASET USING PANDAS. 
variablename = pd.read_csv('iris.csv')

# SHOW THE TOP 5 ROWS OF THE DATASET, WHICH HAS NOW BEEN DEFINED AS A VARIABLE. 
variablename.head()

# SHOW THE BOTTOM 5 ROWS OF THE DATASET. 
variablename.tail()

# (PLT) PLOT BASIC MATPLOTLIB GRAPHS. 
import matplotlib.pyplot as plt
plt.scatter(x,y)   # SCATTER
plt.plot(x,y)      # LINE
plt.hist           # HISTOGRAM. 
plt.figure         # CREATE EMPTY CANVAS FOR MULTIPLE PLOTS ON SAME GRAPH. 

# (PLT) ADD LABELS TO GRAPH AXES AND TITLE TO GRAPH.
plt.xlabel('x label')
plt.ylabel('y label')
plt.title('Title of Graph')

# (PLT) ADD ATTRIBUTES TO LINE PLOTS. 
plt.plot(x,y,'b--')
# b-- indicates colour and line style. 

# (PANDAS) PLOT PANDAS BOXPLOTS.
import pandas as pd
variablename.boxplot

# (SEABORN) PLOT SEABORN BAR GRAPHS.
import seaborn as sns
sns.barplot(data=dataname[['sepal length', 'sepal width', 'petal length', 'petal width']], orient = 'v')

Assignments

1: How will CURATE.AI enter the market?

  • Look at market-approved software (SAMDs and CDSS) which have been passed through the FDA

  • Research press releases, quarterly reports and financial statements (if company is public), news articles to see how they entered the market

  • Help Curate.AI find its pathway to market

2: Market Research report & Presentation to Industry Professionals

  • Write report on multiple myeloma, existing treatments (VCD, VTD), the cost of these treatments (cost to individual, cost to society, cost of side effect management), and the demand for these treatments (prevalence of disease)

  • Give a presentation overview of the market for these drugs, as well as background information on the condition itself