Forum:From protein-protein interaction networks to..

Question

[BioInfoBeginner]

Hi all,

The field I'm learning about has a lot of papers which usually go something like NGS data about a disease into differential expression analysis, enrichment analysis and protein-protein interaction networks resulting in a hub gene/s and proposed biomarker. Looking back on these papers from the past decade some portion seem to have few citations or be elaborated on any further than the original finding.

As a result when reading through the literature it has left me with the feeling of there being an enormous gap between where the current body of knowledge and computational tools are and getting to a point where we understand the biology of diseases to a similar extent that we understand something like the citric acid cycle.

• Have I missed a section of the literature and am getting a biased view or is establishing hub genes and a network the fullest extent that computers can currently contribute and the rest is up to wet-lab innovation?
• Can anyone recommend work of labs who might be attempting to improve this situation?
• Have any of you experienced a similar phenomena with your niche in this field?
• Is it just an unfortunate reality of the logistics of validation that only a smaller number of these findings get investigated in the wet-lab?
• Or have I misunderstood and mischaracterized the state of things?

Just hoping for some discussion so if you have any thoughts, even tangential, please contribute.

Answer 1

[Mensur Dlakic]

I am sure you know this, but it is worth saying it: you are talking about molecules here. Many different molecules. Understanding how invisible things work requires many detailed and painstaking experiments. Let's also keep in mind that many pathways, unlike citric acid cycle (CAC), have more than 20 reaction intermediates that are not necessarily very abundant.

Each of your questions could be a discussion on its own, so I will just offer a general view. It is true that we know more about CAC than other pathways, just like we know more about E. coli than about Nauta Salamanders. Part of it is historical as some pathways/organisms have been studied for a long time, or have been deemed more important to understand. We are far from understanding the biology of all diseases, but we understand some of them. We are also far from understanding biology in general, but you should have seen where we were 20-30 years ago. Anyway, studying macro systems with many moving parts is difficult enough even when we can see and touch things, and measure them directly. It is at least a couple orders of magnitude more difficult when we are trying to understand invisible things such as molecules.

[BioInfoBeginner]

Thanks for chiming in Mensur. Yes, a very good point the degree of complexity is significantly more in our biological problems than my crude comparison.

I should add to, my intention is not to criticise the publishing of papers that end at the hub gene and biomarker point but rather to try and get a feel for how individuals with a lot more experience, such as yourself, think that point might be extended going forward. Thanks for pointing out historical context, I dare say those of us joining the field at this stage are very spoilt with the tools at our disposal compared to 20~30 years ago.

The work of Kotlyar et al seems to be the most ambitious I've found that is optimising all currently available information. Are there others you like the approach of?