Jupyter Notebook project to analyze and create insightful plots of a drug regimen dataset
We investigated the effectiveness of the anti-cancer drug, Capomulin, against numerous other treatment regimens for the treatment of squamous cell carcinoma (SCC), a commonly occurring form of skin cancer. This study involved 249 mouse models (approximately 50/50 male/female ratio) with SCC tumors, 10 treatment regimens, including a Placebo, over the course of 45 days. The number of mice used for each drug regiment was as follows. For the Capomulin and Ramicane 230 and 228 were used respectively. For Ketapril, Naftisol, Zoniferol, Placebo, Stelasyn, Ceftamin, and Infubinol between 178 and 188 mice were employed. For the Propriva regimen, only 148 mice were used.
Over 45 days, the mean tumor volume (mm3) for Capomulin regimen was about 41 mm3 with only the Ramicane regimen showing a similar tumor volume of about 40 mm3. This suggests that Capomulin and may be as effective as Ramican at treating SCC, of course additional studies are needed for further verification. The other drug regimes, including the Placebo, exhibited tumor volumes of between 52 and 55 mm3 likely indicating these drugs were ineffective at treating SCC. Looking at the final tumor volume of a subset of the regimen drugs which included Capomulin, Ramicane, Infubinol and Ceftamin at the 45 day mark more clearly shows the difference in treatment effectiveness of the first two versus latter two drugs. This data is summarized in the attached box-plot that shows Capomulin and Ramicane have almost identical mean tumor volumes of 36 mm3, while the other two drugs have mean tumor volumes of approximately 58 mm3. However, it should be noted that the Infubinol data has a single outlier with a tumor volume value of 36 mm3.
In order to see how the Camuline regime affected the tumor volume over the 45 day study, the data for a single mouse (mouse ID: l509) was plotted over time. The plot shows that there is in fact a decrease in tumor volume at the 45 day mark (~42 mm3) compared to the starting volume (45 mm3). Additionally, a plot of mouse weight (g) vs tumor volume (mm3) was created and it shows a linear correlation (correlation of coefficient value of 0.84) between the two parameters. The heavier the mouse, the larger the average tumor volume and vice versa. However, at this point we cannot say that bigger mice weight is causing larger tumor volumes. Correlation does not always equal causation, and other factors (i.e. age, diet, etc) need to be considered to reach any conclusion.