Identification of drivers’ signature of disease

Introduction

One could define disease signatures as prototypical patterns of entities, that interact together and associate with the disease physio-pathology. In other words, disease signatures are not only made of genes and proteins, as latest advances in disease progression demonstrated involvement of processes, cell types, that eventually support the emergence of clinical signs and symptoms.

• It is for note that drivers’ signature of disease only applies to the entities that contribute to the pathological progression.

A Causal avatar is an interactive map that represents the disease progression steps, from 3 angles: time scale, context that includes hierarchy of events, as well as physiological scale (intracellular to systemic multi-tissue organization). Such avatar of disease progression is really neat to build rigorous knowledge-driven hypotheses of drivers’ signature of disease.

Materials

• Causal Avatar from the Reckonect Library of Disease Avatars (Avatar of Endometriosis for the example),
• Reckonect access to avatar.

Case Study: Endometriosis

Endometriosis is a condition that is characterized by the presence of endometrial tissue outside the uterus. It affects more than 10% of women during their reproductive age, that is to say at least 190 million people worldwide. The etiology of the disease is still being discussed, but the theory of retrograde menstruation, in other words the incomplete clearance of menstrual debris and their migration to ectopic sites, is considered as the most plausible cause (boxes M3/M4 in Figure 1).

Methods & Results

There are two ways to identify potential drivers of the disease physio-pathology. Either you can use the whole avatar, or alternatively, you can focus on a specific transition of the avatar (as the transition 68 on avatar of endometriosis in Figure 1).

Method #1: identify drivers from the whole avatar (Figure 2)

Using the whole avatar, you can find the major potential molecular drivers (from the protein recapitulating table). The proteins are ranked according to their appearance frequency in the entire avatar. It means that the more a protein appears in the boxes, the more it will be at the top of the list. Thus, a protein that is present at the top of the list could be considered as an important putative driver of the physio-pathology.

This hypothesis, has to be confirmed by further experimental investigations.

Method #2: identify drivers by focusing on a specific transition (Figure 3)

Should you want to focus on a given step during pathological evolution, you might want to use a more precise approach on a peculiar edge transition.

For example, if you are interested in the identification of drivers at the heart of the migration of menstrual fragments to ectopic sites, you can focus on boxes M3, M4, and the transition edge number 68.

When you click on each box, you have different entities which describe the box (MESH for biomedical entities, Go-BP for Biological processes, CL for Cell types, and Proteins). When you click on the edge, you’ll have both the literature references and sentences that were used to established the link between the boxes, as well as the proteins that are involved in this transition.

Conclusions

Altogether, our results demonstrate that Causal avatar are very useful to identify putative drivers of disease progression. More precisely, we described 2 complementary methods to unravel molecular leverage to understand pathological evolution, as a whole, or for specific transition.

What’s next?*

You can investigate knowledge borders to formulate new & out-of-the box hypotheses (see Identify prospect to a field of research literature).

You might want to understand how the ballet of the drivers accounting for the pathological progression (see Retrieve mechanism of action from literature).

In case you would be willing to integrate multiple data types (e.g. multi-omics, semantics, imaging annotations), we have your back (see Candidate use of experimental data to consolidate r&d avenues or use the Manta Platform for multi-types data integration).

Finally, you could either be willing to understand the bioavailability of the drivers that you are interested in (blood, urine, feces, tissue specificity) to fuel your R&D for diagnostics solutions.

*Please see the application resources on our Access Hub.