whats the longest experiment you’ve ever tried but it failed?

Matthew Bareford answered on 31 May 2019:

When your sequencing DNA, it takes a total of 3 days to complete each experiment run…. and you only know if it’s worked at the end!… that’s a long time when it doesn’t work!

Rebecca Moon answered on 1 Jun 2019:

All my research is on real live humans so thankfully none of it has ever gone terribly wrong. Children and teenagers come to our research centre with their parents and we do scans, measurements, questionnaires and sometimes take blood samples. The visit can take 2-3 hours some times, and we may see 1000 children for the study, which can take several years to gather all the data…and then the results are not what we expected! That’s always disappointing, but I wouldn’t call it a failed experiments as we can still learn something from it.

Marianne King answered on 1 Jun 2019:

I have to turn stem cells into another type of cell called neural precursor cells (these make almost all of the cells that we have in our brains) and this takes 17 days. This process has gone wrong for me many, many times, often around day 12. But eventually you improve, and it is true that you do learn from failure. Even though it feels awful at the time.

Aina Roca Barcelo answered on 1 Jun 2019:

mmmm… this will sound really awful and discouraging but it was a cell culture that we were growing for 1 month or so to then apply some drugs… but they died!

In case you don’t know, cells are like babies, they need your attention all day/all night, you need to feed them, make sure they are warm enough, that there is no dirt around (they hate dirtiness cause it can make them sick!)… in other words, you become a mum. And I was definitely NOT ready for that! haha So after 1 month of taking care of my cells they decided it wasn’t enough and they sadly died. This meant that I had to start from scratch all over again! Yay!

Good news are, I learned my lesson! My second batch grow beautifully well–she said in a proud mum voice.haha

Kate Timms answered on 1 Jun 2019: last edited 1 Jun 2019 11:57 am

One of my experiments in my last job was over a month long! The cells I used took a whole month to ‘grow up’ and turn into the type of cells I needed (we call this ‘differentiating’).

Once, after a whole month of differentiating the cells, something went wrong on the last day and I had to start all over again. That was… well lets just say I had to go outside and scream before starting again!

Luckily that was the only time it didn’t work.

Kaitlin Wade answered on 1 Jun 2019:

When I was studying Biology as my degree, I remember that I had to work on a special type bacteria that had a gene that turned on and off in an oscillation over the course of the day. We had bought this bacteria that had green fluorescence tagged onto that gene so that whenever the gene was turned on, the bacteria and its surroundings shone green. We were interested in capturing this with photos under the microscope and also model the oscillation time with maths on a computer. So, I had to culture these bacteria on a plate (which took a couple of days to do) then I had to spend the best part of a week within the darkened microscope room to take pictures of these bacteria and the green light they were producing every 5 minutes for hours on end. I must have stayed in the microscope room for about 4 hours each day for about a week until I realised that the microscope camera had been set up on an ‘auto’ exposure, meaning that each picture wasn’t comparable… so… I had to do it all over again until we got some better pictures. Even then, the environment wasn’t helpful for the bacteria gene so we didn’t really manage to capture anything interesting!

I think that’s why I don’t really like lab work now! With epidemiology and computer-based science where the data is already available, my ‘failed experiments’, where I make a silly mistake, take comparatively WAY less time and I can just redo them!

Rachel Hardy answered on 1 Jun 2019:

I would say that the longest one was when I first tried to make adult brain-like cells in the lab. When working with cells for routine experiments, scientists often use cells taken from human tumours. These are more commonly known as cancer cells, and we use them because they divide extremely quickly. This provides us with a large supply for our experiments, which is good as we get results fast. I investigate drugs that are used to treat diseases of the brain, and how we could design them to cause fewer side-effects. Although cancerous brain cells are good for quickly identifying toxic drugs, we need a more realistic model that is more similar to cells in the human brain. Cancer cells contain jumbled DNA which make them very different to normal cells. So, for my work I added a cocktail of chemicals to my cancerous brain cells in a big dish, to turn them into cells more like those found in the adult human brain. This process is called ‘differentiation’. After 3 weeks, this experiment was successful. My cells had clearly changed their appearance, and began to make proteins that are only found in adult brain cells. However, to perform my next experiments on these cells, I had to transfer them from the big dish to extremely tiny plates. This process proved too stressful for the differentiated cells, and caused them to die – such a shame after so much effort had been put into it!. That’s science for you though – and it’s always such an amazing feeling when you do find a way to make the experiment work 🙂 !

Nina Rzechorzek answered on 2 Jun 2019:

Not so much a single experiment but a refinement process. For many of my PhD experiments I needed to grow human brain cells from stem cells; this is quite easy to do now, but at the time there weren’t many labs doing it and when I started, the success rate of keeping the cells happy for long enough to do my experiments wasn’t very high. I spent the best part of a year refining the protocol – effectively learning to ‘listen’ to what the cells needed and when, and optimising a lot of tiny details. It was really worth it because I ended up using cultures that were much healthier, lasted longer, and gave me more reproducible and robust data when it came to the key experiments. Things that I learnt from this process I still use today.

Shobhana Nagraj answered on 2 Jun 2019: last edited 2 Jun 2019 9:04 am

Really good question! I work in an area called clinical research, working with real people, living their daily lives. We usually have an idea of some form of treatment or training, that we think we might help a group of people with a particular problem. This idea is based on reading a lot of evidence around the subject, so you know what is safe and likely to work. We then test whether the idea works by comparing the treatment between groups of people in something called a Clinical Trial. Some clinical trials can take years to complete – like even more than 5 years! It can be really hard if at the end of that time you find that your good idea for a treatment or training didn’t make any difference to people! I guess sometimes, it can be just as important to know what doesn’t work, as it is to know what does work! The best way to make sure you that the results of these long experiments are worthwhile is to read a lot before you start to make sure you have a really good understanding of the area and of the people you are working with, and to ask the right research question, and design the experiment that will best be able to answer your research question. I am currently in the process of trying out an idea that might work, using digital technology (a mobile App) in rural parts of the world – I am hoping that it will make a positive difference to the people involved!

David Wilson answered on 3 Jun 2019:

Some of my experiments can last up to 6 months, I wouldn’t say any of them failed but sometimes you don’t get the result you expected. If you did everything in the experiment correctly then this can be really interesting.

Why didn’t the experiment do what you thought it might?

We can always learn something from a well designed experiment, it’s best not to assume you know exactly what’s going to happen as you can then be open to other results.

Ettie Unwin answered on 3 Jun 2019:

During my PhD I ran computer simulations that took a week to find that I made a small typo in the code, or the maths I wanted the computer to solve, so the results were wrong. Another time a powercut took out the machine I was using and I had to start again.

Anabel Martinez Lyons answered on 3 Jun 2019: last edited 3 Jun 2019 9:46 am

Great question! I have done a technique called CRISPR, which allows me to edit DNA in cells. It takes up to 6 weeks to go from start to finish (I know, a long time!) – you have to design and produce the construct that will do the editing which takes a week or more, then grow the cells that you’ll edit, do the actual editing experiment and ‘sort’ them which takes a few hours using a special machine that looks for a marker that indicates that the editing has been successful and filters ‘good’ cells into a tube for further growth and culturing, and then growing of the cells over a few weeks, purification of DNA from these cells and then amplification of the DNA region of interest and (finally) sequencing of the DNA to check it’s all worked! As you can guess, it sometimes doesn’t work (or frustratingly the DNA has changed but not as intended), and it can be really heartbreaking. But you just try again and eventually it does work (promise)!

Deepak Chandrasekharan answered on 3 Jun 2019:

‘If we knew what we were doing, we wouldn’t call it research.’ – This may have been Einstein who said it but the point is, stuff doesn’t work in the lab all the time. Sometimes it’s because of mistakes you’ve made and sometimes it’s because no one has figured out how to do it yet. The key is to learn from it and push things forward.

I spent weeks growing cells, adding lots of things to them to make them various colours to see how fast they were growing. I then accidentally threw the wrong bit down the sink. Argh!

Next time, I did everything correctly, but it still didn’t work. Part of science is trying to figure out why. The most exciting bit is often not when you get the result you expect but when you find something unexpected and think ‘that’s strange…’ as this is new knowledge!

Alex Blenkinsop answered on 4 Jun 2019:

During my PhD I spent months on a project writing a simulation program to imagine how a very complicated clinical drug trial might happen in practice by making up patients and data and seeing how well they do on the different treatments. In the end I couldn’t check the program was right (because there is no answer sheet for this kind of work) so I had to abandon the project and move on. It was very frustrating but part of research as a scientist is learning from the mistakes and preparing better for them the next time.

Thiloka Ratnaike answered on 4 Jun 2019:

My longest experiment that didn’t work out was one that I completed overnight (around 1 or 2 am in the lab), and then found out that I had forgotten to put in a product that was necessary for the DNA analysis to work the following day! I learnt from this experience that rest is important because the brain doesn’t focus that well when you are overtired (even if you don’t feel it physically)- it can lead to errors in simple tasks!

Lorena Boquete Vilarino answered on 4 Jun 2019:

I once did a 3-month experiment looking at how stem cells made immune cells, I wanted to compare the immune cells they made for a long time. After 2 months and a half my cells got infected with bacteria and died… It was not a great day!

Matthew Burgess answered on 5 Jun 2019:

None of my individual experiments are too long. Once at the end of our 14 day infection model the cells that we then use to test how much virus is present in the lung samples were not growing correctly so I was unable to find out if how we had treated the mice during the infection had had any effect.

I spent most of the first year of my PhD trying to get a yeast 2 hybrid screen to work (quite an old fashioned way now to test if two proteins can interact) and just couldn’t get one of the proteins to express. Eventually we had to give it up as a dead end and try a different project.

Shonna Johnston answered on 8 Jun 2019:

As I work in a facility, most of my own experiments are quite short-term so maybe a few days max. This would be similar to previous jobs I’ve been in too.
However, the samples that researchers bring to the facility can be the result of month to year long experiments. Cell sorting is often a technology needed towards the endpoint so its really important that we get our part right by ensuring the equipment is working and setup well.

Question: whats the longest experiment you've ever tried but it failed?

Comments