Case Studies

The Cancer Genome Atlas

View Graph

Problem:

We use Brainome to identify genes that are consistently over-expressed in different tumor types

The Data:

21,000 gene expressions for patients with 33 different types of cancer and a sample set of healthy patients combined into a 11,000 row, 21,000 column, 34 class table

Result:

Brainome model achieves 90% accuracy in predicting cancer type based on over expression of specific genes

Theoretical physics

View Chart

Problem:

Rework of 2014 model published in Nature Communications:

Searching for Exotic Particles in High-energy Physics with Deep Learning

The Data:

5 million observations; 18 features; 2 classes

Result:

Brainome model has slightly better accuracy than published model

Published model is 150X larger than Brainome model

Brainome model produced in under 2 hours, published model took months

IoT & Mobile

Problem:

Video infrastructure client needs very small models for mobile devices

The Data:

16K instances; 107 features; 2 classes

Result:

Both Brainome and sklearn produce models that achieve 85% accuracy

Brainome’s model is 100 Kilobytes in size whereas the sklearn model is 20 Megabytes (200X larger)

Bank Telemarketing

View Demo Video

Problem:

Brainome builds three models in under 10 minutes for the Bank Telemarketing dataset provided by Satoshida Tamoto, via Kaggle. The classification goal is to predict if the client will subscribe (yes/no) a term deposit (variable y).

The Data:

Result:

Brainome model achieves over 90% accuracy in all three model builds with minuscule file sizes. Read our full demo and case study here.

Microbiome research

Problem:

Identify predictors of inflammatory bowel disease (IBD)

The Data:

1,627 observations; 71,266 features; 2 classes (healthy vs IBD)

Result:

Brainome produces a model in only 15 minutes*

Model achieves 99% TPR for IBD & 94% TPR for Healthy

Model is only 100 Kilobytes in size

200 predictive features identified, ranked by order of importance

*Data set analyzed using Dell XPS-13 laptop, Intel i7 CPU 3 GHz, 32 GB RAM, 2 TB SSD

Genotype-Tissue Expression

Problem:

We use Brainome to classify tissue types based on RNA seq-data

The Data:

16K observations; 56K features (genes); 21 classes

Result:

In 1.5 hours* Brainome produces a model that achieves 98.69% accuracy

Model is only 478 Kilobytes in size

273 predictive features identified, ranked by order of importance

*Run on 48 Core AMD Ryzen 3960X, 128 GB RAM

Interested in our case studies? Let's create one together!

or stay in touch

Case Studies

The Cancer Genome Atlas

Problem:

The Data:

Result:

Theoretical physics

Problem:

The Data:

Result:

IoT & Mobile

Problem:

The Data:

Result:

Bank Telemarketing

Problem:

The Data:

Result:

Microbiome research

Problem:

The Data:

Result:

Genotype-Tissue Expression

Problem:

The Data:

Result:

Interested in our case studies? Let's create one together!

Product

Resources

Company

Legal

Share it