UMAP is superior to t-SNE and opt-SNE in terms of identification of CLL cells
Our pipeline using k-means clustering could output accurate tumor burden estimation for CLL
Constructing a random forest model using the clusters from our pipeline helped sample level diagnostics for both CLL and ALL
https://www.seekpng.com/ipng/u2q8a9q8r5a9u2o0_kawaii-png-download-image-earth-cartoon-png/
In the future, we hope that our pipeline could assist the diagnosis of leukemia and find new markers. For example, consider a doctor in Alaska who treats 1 leukemia patient per month and a doctor in California who treats 10 leukemia patients per day. Their experience on leukemia cases could vary greatly and thus provide different diagnoses and treatment plans for the same case. In this case, doctors can use our pipeline as a validation. Moreover, they can develop more precise treatment plans based on tumor burden estimation.
Page Leader: Ye Jin