Slides online at:



Wednesday October 11, 2017

6:00–9:00 pm, Milkboy Philly




putting machine learning in the hands of cancer biologists

Cognoma Motto 1

The Cancer Genome Atlas

$425,000,000 project funded by the U.S. Government

33 types of cancer

  • esophageal carcinoma
  • glioblastoma multiforme
  • head & neck squamous cell carcinoma
  • kidney chromophobe
  • kidney clear cell carcinoma
  • kidney papillary cell carcinoma
  • acute myeloid leukemia
  • brain lower grade glioma
  • liver hepatocellular carcinoma
  • lung adenocarcinoma
  • lung squamous cell carcinoma
  • mesothelioma
  • ovarian serous cystadenocarcinoma
  • pancreatic adenocarcinoma
  • pheochromocytoma & paraganglioma
  • prostate adenocarcinoma
  • rectum adenocarcinoma
  • sarcoma
  • skin cutaneous melanoma
  • stomach adenocarcinoma
  • testicular germ cell tumor
  • thyroid carcinoma
  • thymoma
  • uterine corpus endometrioid carcinoma
  • uterine carcinosarcoma
  • uveal melanoma
machine learning

Predicting cancer mutations from gene expression

Using public domain TCGA data

(from Xena Browser)

7,306 samples

20,468 genes with expression


21,960 genes with mutation.



elastic net logistic regression

Precision Oncology

"the science of ensuring that a patient's treatment is specifically designed and targeted for their unique form of cancer"

everyone learns something new

Cognoma Motto 2

Tonight makes 27!

Pull requests model for contribution

Contribution per user by GitHub repository

35+ contributors

Childhood Cancer Data Lab


still accepting contributions


Cognoma Launch Party

By Daniel Himmelstein

Cognoma Launch Party

Slides for the Cognoma Launch Party on October 11, 2017 at MilkBoy Philadelphia. Meetup info at Pictures from this event at

  • 2,608