My submission for Stage 1 of HackBio internship.

The full code is included in the simple_code.py file.

For this stage, I generated:

1) Python functions to perform hamming distance between 2 strings(Slack username and X username) or translate protein, both based on user input.

2) Python code to replicate a figure panel based on meta-analysis of human tissue(see below):

Part A – Gene Expression Analysis(panels a-b):

Samples: HBR (Human Brain Reference) and UHR (Universal Human Reference): HBR and UHR consist of gene expression counts from Normal human brain or various cancer tissues, respectively.

Goal: Examining gene expression between normal and cancer tissue.

Part B – Breast Cancer Data Exploration(panels c-f)

Samples: Wisconsin Breast Cancer Database(WDBC): Breast cancer images classified as Malignant(M) or Benign(B) based on 30 numerical features.

Goal: Compare qualitative features between benign and malignant Breast cancer samples.

Figures:

(a) Clustered heatmap of the top differentially expressed genes based on the normalized counts between HBR and UHR. (b) Volcano plot of the differentially expressed genes on chromosome 22 between UHR and HBR samples. (c) Scatter Plot of (radius vs texture), colored by diagnosis (M = malignant, B = benign). (d) Correlation Heatmap- Heatmap of correlation matrix(plus annotation) of six key features: radius_mean, texture_mean, perimeter_mean, area_mean, smoothness_mean, compactness_mean. (e) Scatter Plot of Compactness_mean vs smoothness_mean colored by diagnosis. (f) Kernel density estimates (KDE) of area_mean for Malignant(M) and Benign(B) samples.

Findings:

Part A- Generated many differentially expressed genes that are visibly altered in UHR(Cancer) vs HBR(normal tissue).

Part B- There is a strong correlation between many of the features in the WDBC, especially between perimeter, area, and radius means. A distinct distribution profile was observed between Malignant vs Benign for area mean.