1
00:00:00,840 --> 00:00:05,280
We do very much data-driven work
and I am always very much of the opinion

2
00:00:05,280 --> 00:00:07,600
that we should gather our data
to answer a question.

3
00:00:07,600 --> 00:00:10,320
But that very much doesn't 
mean that's the only 

4
00:00:10,320 --> 00:00:12,480
question that data can answer.

5
00:00:12,480 --> 00:00:15,400
These experiments
costs tens, hundreds of thousands,

6
00:00:15,400 --> 00:00:17,280
sometimes even millions of pounds to execute,

7
00:00:17,280 --> 00:00:21,080
so it's very important
to follow the FAIR principles

8
00:00:21,080 --> 00:00:24,160
in order to make data as accessible
to other people,

9
00:00:24,160 --> 00:00:27,800
to enable other people
to be able to take full advantage

10
00:00:27,800 --> 00:00:31,800
of the investment that patients,
the public and medical charities have made

11
00:00:31,800 --> 00:00:33,920
in generating that data.

12
00:00:37,480 --> 00:00:41,200
In this project, we were interested in a cancer
called multiple myeloma.

13
00:00:41,200 --> 00:00:43,360
It’s a blood cancer

14
00:00:43,360 --> 00:00:47,040
and it is a particularly
heterogeneous cancer.

15
00:00:47,040 --> 00:00:50,160
And what that means is that different
people have very different cancers.

16
00:00:50,160 --> 00:00:53,720
What we’re interested in
is identifying the places

17
00:00:53,720 --> 00:00:58,480
where those pieces of DNA were that were
controlling the expressions of the genes.

18
00:00:58,480 --> 00:01:03,240
So, we are involved
in generating very large datasets

19
00:01:03,240 --> 00:01:04,440
and then using those datasets

20
00:01:04,440 --> 00:01:07,200
to ask questions
about how genes are regulated

21
00:01:07,200 --> 00:01:08,920
and when they're switched 
on and switched off 

22
00:01:08,920 --> 00:01:10,840
and how that goes 
wrong in disease.

23
00:01:10,840 --> 00:01:14,880
The biggest challenge is almost always

24
00:01:14,880 --> 00:01:17,880
convincing people who aren't used

25
00:01:17,880 --> 00:01:21,400
to this way of working, to work this way.

26
00:01:21,400 --> 00:01:26,680
Sometimes it can be difficult
to convince collaborators to release data.

27
00:01:26,680 --> 00:01:30,680
When it came to the end,
I had quite a lot of work to do,

28
00:01:30,680 --> 00:01:34,080
tidying up the code, rewriting some of the code

29
00:01:34,080 --> 00:01:38,520
to get it into a state where it 
could be useful for other people.

30
00:01:38,520 --> 00:01:41,400
I don't think that code has to be perfect
when it's released.

31
00:01:41,400 --> 00:01:45,680
I think we should be encouraging people
to be releasing non-perfect code

32
00:01:45,680 --> 00:01:48,760
because otherwise lots of people
will never release code,

33
00:01:48,760 --> 00:01:52,200
 but it does have to be in a state
where it is useful for others.

34
00:01:52,200 --> 00:01:53,720
It's actually one of the great things

35
00:01:53,720 --> 00:01:57,880
is that we can just go out there
and we can use the data

36
00:01:57,880 --> 00:02:01,440
without having to be in contact
with the producers.

37
00:02:01,440 --> 00:02:06,480
FAIR principles are just completely
indispensable for the work we do.

38
00:02:06,480 --> 00:02:09,000
The data is so expensive to collect

39
00:02:09,000 --> 00:02:14,320
and there is so much information
in the datasets we collect

40
00:02:14,320 --> 00:02:17,600
that are not relevant
necessarily to the question we're asking,

41
00:02:17,600 --> 00:02:22,920
that it's just inconceivable that this data 
shouldn't be out there being reused.

42
00:02:22,920 --> 00:02:25,560
There's nothing more frustrating than reading a study

43
00:02:25,560 --> 00:02:29,360
and finding that you can't make use
of the data from that study,

44
00:02:29,360 --> 00:02:34,000
or that you can't interpret it because
you can't get the details of the analysis.

45
00:02:34,000 --> 00:02:38,680
So, I'd urge people to very much push for
and maybe even help

46
00:02:38,680 --> 00:02:44,440
create discipline-specific repositories
that have defined metadatas

47
00:02:44,440 --> 00:02:48,040
and ways of describing metadata 
and ways of storing the data.

48
00:02:48,040 --> 00:02:50,000
As researchers,

49
00:02:50,000 --> 00:02:54,120
our aim is to bring knowledge
into the world,

50
00:02:54,120 --> 00:02:58,520
and FAIR is a way in which we do that.

51
00:02:58,520 --> 00:03:00,760
We allow that information to get out there

52
00:03:00,760 --> 00:03:04,200
and that information to be findable
and the work that you've spent.

53
00:03:04,200 --> 00:03:09,000
You’ve invested so much time in that
to make the difference that it can make

54
00:03:09,000 --> 00:03:12,080
rather than just sitting on a dusty shelf
somewhere.

55
00:03:12,080 --> 00:03:15,840
It helps us to achieve
what should be our ultimate goal

56
00:03:15,840 --> 00:03:21,480
as researchers in changing the world
and increasing the sum of human knowledge.