• Question: Are your conclusions based on a lot of accurate data and if so, what is this data?

    Asked by juniorgujral to Lyn, Katy, Paul, PB, Ruth on 14 Jun 2013. This question was also asked by shaanzafar.
    • Photo: Peter Balfe

      Peter Balfe answered on 14 Jun 2013:

      One of the hardest questions you can ask is “what do you mean by accurate?”!

      I like the fact that much of my work is based on gene sequences. There the data is as solid as it can be, the sequence is A or G or C or T, like it or lump it! No interpretation, there it is.

      But most biological data is more nuanced, reflecting the fact that organisms are complex. What is true for an organism in one condition may be false if you change things.

      For example the famous line “the doubling time of E. coli is 20 minutes”: Well yes, if you give it ideal growth conditions and grow it alone! But in real life it will be in a mix with thousands of other bacteria and fighting it out for scarce resources, so that number simply won’t apply.

    • Photo: Paul Waines

      Paul Waines answered on 14 Jun 2013:

      my conclusions have to be based on accurate data, otherwise it is often hard to come to a conclusion with any confidence. Sometimes you have to be prepared to talk about why you have come to your conclusions to the boss!

      This data might take various forms – microscope images, graphs, tables etc. – and is quite often investigated using statistics (using maths to look at large sets of data). Although I don’t enjoy this part too much, it helps you to make sense of what all that data means!

      If I am looking at pictures taken using a microscope, it is often a case of putting forward my own ideas as to what they show- I like this part, because there is no maths involved!

    • Photo: Ee Lyn Lim

      Ee Lyn Lim answered on 14 Jun 2013:

      Like Peter said – it’s hard to say what ‘accurate’ means! When we say something like a news report is inaccurate, that’s easy to tell because we can quite easily find out what actually happened, and how much of that report is correct. Experimental results are a bit like a news report about an invisible world – there’s no way to ask an eye-witness if it’s all true!

      So we try to get data from as many samples as possible (if a hundred thousand cells all say they have Protein A but not Protein B, there must be some truth to it…). My work involves running millions of cells through a laser beam and looking at what colours of fluorescence they give off. No two cells give exactly the same profile, so I end up looking at a lot of graphs, where cells that look similar are clustered together. Then it’s up to statistics to tell me if two clusters are far enough apart to be considered two separate types of cells, or if it’s just a random distribution. And statistics often gives me an answer that sounds like this – ‘There is a 95% chance that these are different types of cells, and only a 5% chance that they’re actually the same type of cell but just happen to look different.’ If that’s the case, I’ll make a decision that my results are accurate, and treat the cells as two different types from that point on.

    • Photo: Katy Brown

      Katy Brown answered on 14 Jun 2013:

      My conclusions are based on whatever the most accurate data available at the time is. For lots of species of animal, reference genome sequences are available online – these are sequences showing the best estimate so far, based on the work of lots of different scientists, of the whole DNA sequence of that animal. This data will never be 100% correct, but it is regularly updated to be as accurate as possible. I work with these sequences and compare them to each other.

    • Photo: Ruth Mitchell

      Ruth Mitchell answered on 17 Jun 2013:

      I agree with all that is said by the previous scientists.

      Think about if our conclusions weren’t based on accurate data: what we do will have future implications on how diseases are treated and if we tell people lies then they might be treated wrongly!! That would be terrible!

      Like Lyn said, we work on living things so each one looks a little bit different, so we have to collect a lots of data(each experiment must be done 3 times for the scientific world to believe it) be sure that what we are seeing is true and not a one off phenomenon. And then it’s up to maths to tell us if it is true!

Comments