Forum:what is next going to happen in bioinformatics?

Question

[Mo]

In fact many people around the world are working in this domain. Some studied bioinformatics and some not (even I see physician are doing bioinformatics). I have been reading papers from all known journals which publish biology related bioinformatics papers or pure bioinformatics. I can tell, pretty much around a topic all times. I know it is very general question and we cannot give a great and direct answer to it. However, I would like to know which topics you think are the hot spot these days for bioinformatics?

For example, many people are doing sequencing (of course we cannot have a golden standard because "all modelling are wrong but some are useful") so these types of studies are going to be forever?

We all know that bioinformatics is only a tool and not the pure science itself. so can we think that it is a died field since mathematics/statistics found itself already or so much left to do? If so much left to do, what could be those topics?

I am so eager to know about your opinion

[karl.stamm]

Go back to the BioStars frontpage and read the questions that don't say "(Closed)". That's what the field is up to.

[Alex Reynolds]

Visualization development and improvement will be important, if only because humans are mostly visual and this field generates so much data, which needs intelligent reduction.

[Mo]

@Alex Reynolds I definitely agree as this one is always important, however, this has always been the centre of attention for any type of data analysis!

Answer 1

[francis]

One can have several answers, and definitely, the questions on this board is a good set of answers ... Here are some, inspired by many years, many other people's blogs, articles, and top-10 lists, and biased by my own wonderings in cancer genomics in recent years:

1. Open source software, open data portals, open access publication repositories are not only the goal for all, but the norm in all universities, all companies and and all governments, worldwide.
2. De novo assembly from <insert what ever technology, or mix of technologies> that are accurate to 99.99999% and generate the full phased chromosomes set.
3. De novo assembly of all transcripts (long/short/linc etc) to 99.99999% accuracy and for a small number (one?) of cells.
4. Methods that generate all of the variants (sequence and structural) for all tumour types for a given cell.
5. Methods that clinicians can use to prescribe treatment given a sequenced genome, as a push-button operation that all want to use world-wide.
6. Ability to deduce function(s) from structure of proteins.
7. Visualizations tools (as mentioned above) to see structures (all of them), interactions (not hair balls, but useful ways to see networks, pathways and connectivity of a whole cell), multiple sequence alignments (from 1 million genomes) where we can deduce changes, events and interpretations.
8. Somebody to figure out how the environment affect the genome/epigenome.
9. While doing the one above, also figure out how genes are regulated too.
10. Figure out great ways to teach all of this to all, young, old, high school, clinicians, business people, and our very best scientists.

There are probably many more things we need to think about!

[Mo]

@francis thanks! I liked your answer and I am happy to see scientist like you who are not waiting to see a "closed" flag!

Answer 2

[Mary]

If I was a young student right now, I'd head towards the pan-genome analysis and display stuff. How to manage, visualize, and query the huge amounts of individual human sequences we are about to get.

If I was betting on a direction that was going to pay off soon in the field, I think tumor sequence analysis, combined with treatments that are like cocktails--with the appropriate drugs for the right mutations--will be the first big success in bioinformatics.

[Mo]

@Mary thanks for the comment but you are still young (heart is more important than the numbers :-) )

Answer 3

[Jean-Karim Heriche]

From my perspective, bioinformatics is going to include dealing with all kinds of biological data beyond sequences. For example, just look at how light microscopy is now generating data volumes as large as sequencing does and at how quantitative it is becoming. Sequences were just the first kind of biological data to be produced at scales requiring computational handling.

[francis]

Yes, images is indeed a very big one I forgot to mention. Histology, in situ, pathology, histochemistry ... all of these require images that will need lots of processing, counting, aggregation, clustering etc ... yes, will be very important for phenotyping (another one I forgot to mention).

[Mo]

@Jean-Karim Heriche You are right , if we look at the nobel prizes of this year for physics and chem, we surely see the importance of optics toward biology!

Answer 4

[rbagnall]

Accurate assignment of pathogenicity/clinical relevance to non-protein-coding sequence variants.

Application of biosensor tattoo data collected via smartphones.