How is Obesity related to Breast Cancer ? Insights from 140,000 articles.

Yiannis Kiachopoulos
published on August 28, 2018

Obesity is a big problem as more than 70% of American adults are overweight. As we were reading up on the topic we found an article on the website of the American Cancer Society that links obesity to the risk of developing Breast Cancer. The article states, among other things, that women who are overweight tend to have higher insulin levels which could be related to some cancers, including Breast Cancer. This made us curious whether that claim can be reconstructed using the Causaly Knowledge Graph that was created by machine-reading more than 19 Million academic publications in Biomedicine.

To get started and better understand the context of Breast Cancer itself, we took a look into its environment by asking:

What are the Causes of Causes of Breast Cancer with the most evidence in the Knowledge Graph ?

This query returned 300 out of 1,000,000+ relationships. Although this covers only a fraction of all evidence, the resulting diagram reminds us that biological systems are quite complex by nature.

breast_cancer2

This gives us a first view on the environment of Breast Cancer and we can take note of the following:

  1. There are entities that interact directly with Breast Cancer, i.e. where the distance from (A)->(B) is zero such as (Glucose)->(Breast Cancer).
  2. Entities are of different categories i.e. Glucose is a Biologically Active Substance while Recombinant Insulin is a Hormone. There are more than 140 different categories of entities that we use to classify all nodes in the Graph - more on that in another blog post
  3. The pathway we are looking for (Obesity)-->(Recombinant Insulin)-->(Breast Cancer) is marked in red.

Since our interest is in the insulin-mediated pathway, we make a second search in the Knowledge Graph which will only show Hormones as mediators:
(Obesity)-->(Hormones)-->(Breast Cancer).

The underlying query machine-reads 143,548 articles within < 2 seconds and returns 53 hormones as potential mediators for the relationship in question.

Wider connecting lines indicate that there is more evidence that establishes the relationship between two entities e.g. (Obesity)-->(Recombinant Insulin) is established in 564 academic publications.

Breast_cancer3

From the diagram above we can see that Insulin mediates the relationship between Obesity and Breast Cancer as described in the article of the American Cancer Society.
Where do we go from here ? There are two ways to continue our research:

  1. Deep Dives: To really understand the nature of the relationship between two entities, we need to selectively deep dive into the evidence from the academic publications. For example, is the underlying evidence based on a Randomized Control Trial or an Observational study ? Is the evidence based on hypothetical statements or definitive statements ?
  2. Increase Level of Detail. We asked for (Obesity)->(Insulin)->(Breast Cancer) but we can further increase the resolution of our search query and ask for (Obesity)->(Entity1)->(Insulin)->(Entity2)->(Breast Cancer) to get more detail on the mechanism of action.

Our Rapid-Search Platform allows users to go all the way from big picture graphs to the textual evidence of the machine-read publication. We will explore these options in future blog posts. Stay curious :-) !

Not all evidence is created equal: Machine-Reading in Biomedicine
technology

Not all evidence is created equal: Machine-Reading in Biomedicine

Teaching computers how to read and understand biomedical publications for cause and effect relationships is a challenging task. This is especially true for it might not be intuitive what we mean by "read" and "understand".

Knowledge emergence - what we learn from 100K monthly publications
Point of View

Knowledge emergence - what we learn from 100K monthly publications

Every month we process more than 100,000 scientific documents. New knowledge is emerging every month across thousands of scientific disciplines

How to read 35,000 articles in 1 minute
use case

How to read 35,000 articles in 1 minute

Rapid search for causes of anxiety at super human speed using the Causaly machine-reading platform

Never miss an update

Subscribe to our newsletter