Forum:Have we failed as bioinformatician in this time of COVID-19

Question

[heididunst]

It about more than 5 months for COVID-19 pandemic but still we far from developing any vaccine. However, to date numerous researcher have developed n number of bioinformatics tools, techniques and models claiming these tools will help in early detection and prevention of any disease outbreak. It very strange that in this time of pandemic nothing seems working well. My question is then what is the use of these techniques when we are failing to do so. Or these tools are just for mere publication and promotion.

Additionally, instead of using already established technqiues, many researchers are using single cell techniques to identify key genes associated with COVID-19. Is this waste of money.

I feel like instead of giving fund to tools and techniques, more focus should be given to real wet lab research in order to combat this deadly disease. Note, I am also a bioinformatician. But as a bioinformatician, I feel we bioinformatician have failed. There is no point in giving bioinformatics as a special status.

Rest in peace to all bioinformatics tools and techniques :(

[igor]

I feel we bioinformatician have failed. There is no point in giving bioinformatics as a special status.

Who has given bioinformatics a special status? I don't think I've heard anyone say that bioinformaticians will find a cure. If you look at the total effort in finding a cure or a vaccine, bioinformatics is a small fraction.

[rprog008]

I totally agree with ignor. Contribution of bioinformatician is minimum toward vaccine development

[piyushjo]

We might have failed as "human beings", but not as bioinformaticians :D. As a person who has experience both in wet and dry lab, without current bioinformatic approaches, we won't be where we are now. The rate at which the genome has been decoded and reasons for the virus increased specificity for ACE2 were understood, would have involved multiple phd years with just wet lab work.

[Alex Reynolds]

Advances in software tools and sequencing technology made in the last two decades allowed bioinformaticists and wet lab scientists to sequence the virus very quickly, and resequence it as it spreads and mutates. This is all useful data for public health experts to use to track infections and guide policy, reducing deaths, and giving more time to those making vaccines and therapies. We're also the ones doing the programming to run ML-guided or parallelized searches for existing drugs that might be repurposed to fight the virus or to lessen symptoms caused by the resulting disease, and to analyse data that come back from target and other studies. If you feel like you failed, maybe pick a different field, but it seems like we've been doing our part in various ways.

[rprog008]

What about evolutionary biologist of elite insitutes. These people claiming of designing so many hypothesis nd bioinformatics tools but still failing to detect the origin of covid19. In this case, already many sequences are there. Yes, few studies have stated that this virus is evolving under purifying selection but few are claiming the sequences generated is wrong anf suggesting error.i am still looking for the link, if i find i will share here. Why these evolitionary hypothesis ☹️

[Joe]

They haven't 'failed' to propose the origins of the virus. There are a couple of competing hypotheses as far as I'm aware (blah blah bats, blah blah pangolins, blah blah snakes)...

The problem is, it's impossible to know the ground truth. Even the person you think is patient zero, is only the first person you confirmed the disease in. It's highly unlikely that that person is the first person the virus was able to infect.

Evolutionary biology is only ever, at best, hypotheses, unless you can obtain samples way back in time.

I'm not really sure what the debate is here? Are you suggesting that evolutionary biologists just have no idea what they're doing and thus we've 'failed'?

[Michael]

Yeah, and what about "her emails"? Sorry, but in what way is this even a scientific discussion?

[Michael]

Look at this: https://www.arcgis.com/apps/opsdashboard/index.html#/bda7594740fd40299423467b48e9ecf6 Then come back and tell me again who has failed in this situation.

[rprog008]

Are u aware that most of the tools required for vaccine development are not working properly. What is the use of developing so many tools. Why not to maintain only few fixed servers nd institute for the same. Some of these database become dead after few month of publication

[Joe]

That's because, if anything, the academic grant ecosystem toward maintenance of software has failed.

How are you supposed to support software when you can't keep developers employed forever?

Even the highly popular Phyre2 structure prediction server almost disappeared last year because the funding bodies refused to renew it's grant, even after thousands of letters of support from the community. It is only still running (IIRC) because it secured a specific CZI grant.

If we can't keep something that popular running, we have no hope of maintaining the more niche/less well known tools.

Answer 1

[Jean-Karim Heriche]

The reality is that bioinformatics is only a subfield of biology and most biological questions still need to be tackled from an experimental point of view and gathering good experimental/biological data is still the hardest part. In my experience, good data analysis (to reach solid conclusions) primarily rests on good data. No amount of tooling will turn crappy data into good data.

Also bioinformatics has failed before as it's been oversold (e.g. let's sequence the human genome and we'll cure human diseases, let's study genome-wide gene expression and will understand protein function, let's do single cell sequencing and we will cure cancer), just like AI is currently overpromising and underdelivering.

As as far as bioinformatics predictions go, the sad truth is that few people trust them enough to invest resources into experimentally testing them because many realize the flimsy nature of the data they are built on.

What I see as the main issue with science during this pandemic is that good scientific principles are going out the window under the pretext there's no time to waste. In my view a lot of what's published now is wasted time and resources because many studies lack proper control and/or are underpowered. Same will happen to reach the correct conclusion by chance but resources will then have to be devoted to disprove a lot of the wrong ones on the way to validating the good ones. One could argue this is science as usual but if time and resources are in short supply, why not invest in doing things right the first time around and reduce the amount of noise we will have to sift through later.

One way to get better studies would be for people to share data and resources instead of each group/country/institute only using the little bits of data and wet lab work they have access to but larger collaborations take times and efforts. But most importantly why not simply do science just as it should be done e.g. why do a screen with no replicates and/or no proper controls when this wouldn't normally be considered good enough?

[Mensur Dlakic]

But most importantly why not simply do science just as it should be done e.g. why do a screen with no replicates and/or no proper controls when this wouldn't normally be considered good enough?

I can see at least two reasons why we are doing less than solid science: altruism and ambition. In general, in times of pressure we do wrong things both for good and bad reasons.

[lieven.sterck]

is that a 'we' as in bioinformaticians or a 'we' as in scientists in general?

[Mensur Dlakic]

We as human race. This also happens outside of science.

[Joe]

I think 3 reasons, the third being simply the practicalities.

It's hard to do rigorous science under normal circumstances (maybe you have precious and limited samples so you can't reach an ideal power threshold for instance), so it's even harder when labs are shut, suppliers are strained etc.

My own biotech startup has been under immense pressure during this period as literally everything takes twice as long, we've lost access to key resources or suppliers etc.

Answer 2

[venu]

Hi Heididunst, my few points

It about more than 5 months for COVID-19 pandemic but still we far from developing any vaccine

If we look at time needed for vaccine development (for previous diseases), one might understand it's not possible to get something reasonable in the span of 1 year, let alone 5 months

However, to date numerous researcher have developed n number of bioinformatics tools, techniques and models claiming these tools will help in early detection and prevention of any disease outbreak.

IMO, the ultimate goal of any bionformatics tool is not early prediction/prevention. It only helps you on your way, with other tools and knowledge to reach a potential solution (in this case prediction). BTW, could you may be give an example which tool claimed it will detection and prevent a disease outbreak? Curious to see.

It very strange that in this time of pandemic nothing seems working well. My question is then what is the use of these techniques when we are failing to do so. Or these tools are just for mere publication and promotion

I agree nothing seems to be working well, that's because we were not familiar with what virus is doing inside a human body. I don't see a tool working very well in normal case failed all of a sudden in case of a pandemic. Also, the field of bioinformatics is not only to help in situations like pandemics, but there are 1000s of diseases humans are dealing with. I am sure bioinformatics/tools played a good role here and there. Don't you think so?

Re single cell techniques

Don't you think single-cell techs are well established? They are producing very valuable insights for other diseases e.g. cancers, alzheimer's disease etc. IMO, using single-cell techs will give far more insights than 'established techniques', for example, any specific cell population that changed their transcriptome after infection, what important genes we can identify to potentially target?

Re funding

The current funding agencies already underestimates the value of bioinformatics/software tools in advancing biology. IMO, wet-lab & dry-lab are equally important. However, I wouldn't call wet-lab as real and bioinformatics otherwise, both are real and sciences.

But as a bioinformatician, I feel we bioinformatician have failed. There is no point in giving bioinformatics as a special status

If yet all, I would blame govt for not taking enough measures in handling situations like this. Many a times, experts advised govt what to do to prevent pandemics based on previous experiences, nevertheless, they have done nothing (can't remember the source of a simulation study performed in a US state in 1950s or in that time).

There are many points, than stated, we should consider before we say 'Rest in peace to all bioinformatics tools and techniques'

Answer 3

[lieven.sterck]

As a bioinformatician, i have screened some drug using computational appraoch (published). However, still that need to be validated by wet lab and I am waiting for that approval. After waiting for at least two weeks, I feel like bioinformatician can never compete with wet lab people what ever the situation it may be.

as said in other replies/answers here already, there is no need to compete with each other! They both are two different fields that complement each other (and the complementation is the key thing here). Yes, it can be frustrating to have to wait for wet lab results (or that they have to validate your findings first) but each has their merits.

Moreover (or why not look at this this way: ) , imagine how long it would take if they have to screen all potential drugs or such. Bioinformatics is then very useful (if not necessary!) to do a pre-screening of things and boil down the list to manageable proportions for a wet lab to validate them. So, indeed, wetlab would eventually get to the/a solution and bioinformatics alone might not, but it's the interplay of both that really speeds things up!

everyday example: I don't know a single wetlab group in my whole institute that would start tackling a gene list of thousands potential candidate genes (let alone a whole proteome), so they are very grateful if bioinformatics can reduce that list to a few dozen (eg through DEG analysis or such)

Answer 4

[Mensur Dlakic]

I will share my thoughts about failure, without addressing your suggestion that money could or should be allocated differently. To me, failure means not succeeding in a defined task, where all the goals are clear and the time-frame is known and tangible. It means not succeeding even when it is known that a given approach works.

I don't think we have a defined time-frame here, or that we know for sure bioinformatics works in this case. Frankly, I don't think we can afford to call any approach a failure since we have not given this enough time, even though it is likely that most of them will not work for one reason or another.

[heididunst]

Yes, i do agree with you. But i still have a doubt how these tools and models are helping us in COVID-19. Only one bioinformatic tool which seems appropriate in the present scenario is BLAST. Other tools seems story telling. This is really very disappointing for being a bioinformatician. Only one thing that coming into my mind, what is the contribution of a bioinformatician in this pandemic. Is bioinformatic really needed. What is the use of using this single cell and other high throughput technologies.

As a bioinformatician, i have screened some drug using computational appraoch (published). However, still that need to be validated by wet lab and I am waiting for that approval. After waiting for at least two weeks, I feel like bioinformatician can never compete with wet lab people what ever the situation it may be.

What about expertise from evolutionlary biologist from premier institute. why they are unable to detect the origin of covid-19. or there protocol and hypothesis also not working.

If these people are unable to detect this basic information after 6 months also, what is the point of claiming so much hypothesis. Other may oppose what I am saying and I am happy to discuss all things here.

Answer 5

[JC]

It about more than 5 months for COVID-19 pandemic but still we far from developing any vaccine.

What are you talking about, maybe you are years to design/produce a new vaccine, pharma has started since day 1 of the epidemic.

However, to date numerous researchers have developed n number of bioinformatics tools, techniques and models claiming these tools will help in early detection and prevention of any disease outbreak.

Yes, I also saw a lot of Data "Scientist", trying to do something, but after they realize they can't, start blaming others

It very strange that in this time of pandemic nothing seems working well. My question is then what is the use of these techniques when we are failing to do so. Or these tools are just for mere publication and promotion.

Many things are working, and yes, many things are just promotion

Additionally, instead of using already established technqiues, many researchers are using single cell techniques to identify key genes associated with COVID-19. Is this waste of money.

hmmm, no diagnosis is still using PCR, qPCR or other well-known techniques

I feel like instead of giving funds to tools and techniques, more focus should be given to real wet lab research in order to combat this deadly disease. Note,

That is why we have grant reviewers, no government is giving money to anyone, there is a protocol to ask for money, review, and assign.

I am also a bioinformatician. But as a bioinformatician, I feel we bioinformaticians have failed. There is no point in giving bioinformatics as a special status.

After, this, please burn your special Bioinformatics ID

Answer 6

[rprog008]

Heidi at some point i agree with you but we cannot say all bioinformatics tools are useless.

Regrading single cell, all i can say it work well in case of other diseases where most of the information is known. However it may not work in case of Covid19 because we dont know how covid19 is working. Another problem associated with covid 19 is sequence generated. Some researcher claim that sequence generated is problematic and contain error so high chances are there that whatever we work and publish, nothing will work in patients.

Yes i too agree with you and always support that wet lab approaches are more powerful and robust then bioinformatics approaches. However few bioinformatician accept this fact. More funding should be allocated when project involve both wet lab and dry lab. Allocating fund only for developing tools nd techniques is not worth investing. These tools are mostly developed in free time of a researcher. So why to invest so much money. If we are investing the same money in wet lab, might be that will be more useful.

Venu, i dont agree that government should be blamed. I belive Government is giving sufficient fund for bioinformatic research and if any tool fail bioinformatician should be blamed. For instance, currently our lab faced so many problem while analysing data with alevin. When contacted original authors failed to solve the problem. So we switched to another tools. Sometimes suerat packages are archive. Function of seurat is changing every now and then. Most of the bioinformatics tools are not managed. Developer develop just for the sake of publication. They dont think about the user. My question is why to update these tools every day. Why not to update these toole after every one year. You see most of the database develop by bioinformatician become obselete after few days. Their existance gone. So why to invest money in developing these databases. Recently orthomam database found to contain error. Reason may be anything but when govt is giving you fund. Why these people are not working effectively.