1
00:00:00,000 --> 00:00:02,670
All right. Our next participant,

2
00:00:02,670 --> 00:00:05,925
is Chastyn Smith from
integrative life sciences.

3
00:00:05,925 --> 00:00:08,580
The title of the talk is
"Helping solve whodunnit:

4
00:00:08,580 --> 00:00:10,380
determining the number
of contributors in

5
00:00:10,380 --> 00:00:13,600
a sample earlier in
the forensic workflow."

6
00:00:13,970 --> 00:00:16,425
We've all probably seen

7
00:00:16,425 --> 00:00:18,510
a crime mystery
show at some point,

8
00:00:18,510 --> 00:00:21,015
whether it was Scooby
Doo or true crime.

9
00:00:21,015 --> 00:00:22,695
So let's play "whodunnit."

10
00:00:22,695 --> 00:00:24,330
We arrive at a crime scene

11
00:00:24,330 --> 00:00:25,890
and we see this puddle of blood.

12
00:00:25,890 --> 00:00:27,375
We take a sample of the blood

13
00:00:27,375 --> 00:00:29,280
and we send it to
the lab for testing.

14
00:00:29,280 --> 00:00:32,505
In the lab, the sample is taken
through the old workflow,

15
00:00:32,505 --> 00:00:35,190
only reveal at the end
that it contains DNA

16
00:00:35,190 --> 00:00:38,010
from maybe one,
maybe two people.

17
00:00:38,010 --> 00:00:39,840
We can't be quite sure because

18
00:00:39,840 --> 00:00:43,295
all the DNA  is not crossing
the level of detection.

19
00:00:43,295 --> 00:00:44,990
Now we could reanalyze

20
00:00:44,990 --> 00:00:47,255
this sample and hope
for better results.

21
00:00:47,255 --> 00:00:49,295
But this process
already took weeks,

22
00:00:49,295 --> 00:00:51,155
and at most it
could take months.

23
00:00:51,155 --> 00:00:54,275
This is a problem in the
forensic DNA community.

24
00:00:54,275 --> 00:00:56,270
Samples that contain DNA from

25
00:00:56,270 --> 00:00:58,655
multiple individuals,
called mixtures,

26
00:00:58,655 --> 00:01:01,895
produce complex results
like the picture on top,

27
00:01:01,895 --> 00:01:05,240
and are not revealed until
the end of the workflow,

28
00:01:05,240 --> 00:01:07,310
at which point it's
too late to do things

29
00:01:07,310 --> 00:01:09,875
differently in order
to get better results.

30
00:01:09,875 --> 00:01:12,530
Now, a solution would be to

31
00:01:12,530 --> 00:01:13,700
detect that a sample is a

32
00:01:13,700 --> 00:01:15,800
mixture earlier in the workflow

33
00:01:15,800 --> 00:01:18,050
so that we could then know to

34
00:01:18,050 --> 00:01:21,050
modify how we
process that sample,

35
00:01:21,050 --> 00:01:24,590
that we can ensure we get
better results later on,

36
00:01:24,590 --> 00:01:26,645
like the sample on the bottom.

37
00:01:26,645 --> 00:01:29,765
Well, this is what
my research is.

38
00:01:29,765 --> 00:01:33,170
I am developing a tool that
can be used earlier in

39
00:01:33,170 --> 00:01:34,790
the forensic workflow that will

40
00:01:34,790 --> 00:01:37,520
determine if a
sample is a mixture.

41
00:01:37,520 --> 00:01:40,340
The way this tool
works is we look at

42
00:01:40,340 --> 00:01:41,360
two different regions of

43
00:01:41,360 --> 00:01:43,969
the DNA that differ
between individuals.

44
00:01:43,969 --> 00:01:45,785
The more differences
that are seen,

45
00:01:45,785 --> 00:01:47,660
the more likely that
sample contains

46
00:01:47,660 --> 00:01:49,415
DNA from multiple people,

47
00:01:49,415 --> 00:01:52,235
and the tool predicts
it to be a mixture.

48
00:01:52,235 --> 00:01:54,740
And the good thing about
this tool is that it is

49
00:01:54,740 --> 00:01:57,455
combined with a step that's
already in the workflow,

50
00:01:57,455 --> 00:02:00,420
so it doesn't add an
increase in time.

51
00:02:00,850 --> 00:02:03,125
So far,

52
00:02:03,125 --> 00:02:05,210
in proof of concept testing,

53
00:02:05,210 --> 00:02:07,669
this tool has been
able to accurately

54
00:02:07,669 --> 00:02:11,015
identify 90% of the
mixtures tested.

55
00:02:11,015 --> 00:02:13,835
Further research
will be done to one,

56
00:02:13,835 --> 00:02:15,665
increase the accuracy
of this tool,

57
00:02:15,665 --> 00:02:17,630
and to develop more features.

58
00:02:17,630 --> 00:02:20,210
For example, we not
only want this tool to

59
00:02:20,210 --> 00:02:23,180
tell you that a sample contains
DNA from multiple people,

60
00:02:23,180 --> 00:02:25,595
but exactly how many people.

61
00:02:25,595 --> 00:02:27,590
The most immediate impact of

62
00:02:27,590 --> 00:02:29,150
this research will be seen in

63
00:02:29,150 --> 00:02:32,795
DNA labs as it will allow for
more samples to be tested,

64
00:02:32,795 --> 00:02:35,210
which could be the
determining factor in a case.

65
00:02:35,210 --> 00:02:37,325
But looking at the
bigger picture,

66
00:02:37,325 --> 00:02:39,170
this research is significant to

67
00:02:39,170 --> 00:02:41,225
society because it is bettering

68
00:02:41,225 --> 00:02:44,270
the DNA analysis
process by allowing

69
00:02:44,270 --> 00:02:46,430
stronger results be obtained

70
00:02:46,430 --> 00:02:49,295
the first time a
sample is analyzed.

71
00:02:49,295 --> 00:02:52,129
This will allow cases
to be solved faster,

72
00:02:52,129 --> 00:02:55,340
the majority of which are
committed by repeat offenders.

73
00:02:55,340 --> 00:02:57,110
So if we can stop a

74
00:02:57,110 --> 00:02:59,690
re-offender before
they offend again,

75
00:02:59,690 --> 00:03:01,340
this will ultimately lead to

76
00:03:01,340 --> 00:03:04,530
a decrease in crime. Thank you.