1
00:00:00,000 --> 00:00:05,880
So we're going to start this talk with a short outline, so we're going to have a biology review

2
00:00:05,880 --> 00:00:10,400
to make sure everyone's on the same page, don't want to leave anybody behind, followed by what

3
00:00:10,400 --> 00:00:15,740
the hell is the second gender code, a little bit of tutorial on information theory, how we use

4
00:00:15,740 --> 00:00:20,500
information theory to hack the second gender code, and then some potential applications. So

5
00:00:21,460 --> 00:00:26,020
since this talk's a little biology heavy, maybe for a biohacking village, we'll have this review.

6
00:00:26,020 --> 00:00:34,260
So I'm guessing most people here know that cells have a genome which contain genes on chromosomes,

7
00:00:34,520 --> 00:00:39,300
and these genes are made out of ATCs and Gs. And this is essentially the generic code for all the

8
00:00:39,300 --> 00:00:43,860
proteins encoded in the cell and all functions encoded in the cell. So how do they actually,

9
00:00:43,940 --> 00:00:50,000
you know, read this and go through it? We have transcription, which reads a gene from the DNA

10
00:00:50,000 --> 00:00:56,000
strand and makes a copy of it in RNA, mainly messenger RNA. And this messenger RNA goes to

11
00:00:56,020 --> 00:01:02,840
the ribosome, and the ribosome is actually the transfer point between the nucleotides, the genome,

12
00:01:03,000 --> 00:01:10,340
and actual proteins. So the ribosome will read the mRNA and bring in tRNAs, and the tRNAs recognize

13
00:01:10,340 --> 00:01:19,760
the mRNA through complementary triplets, so A to U, G to C, and each one of those tRNAs carry an

14
00:01:19,760 --> 00:01:25,820
amino acid and add it to the growing enzyme chain. So that's our basic review. And then

15
00:01:25,820 --> 00:01:26,000
I think we're going to have a little bit more detail on the biohacking process. So we're going to

16
00:01:26,020 --> 00:01:29,380
Most people from high school biology remember something like a translation table,

17
00:01:29,980 --> 00:01:30,940
the genetic code table.

18
00:01:31,300 --> 00:01:36,560
And what's interesting to focus here is that over about, what, 3 billion years of evolution,

19
00:01:36,860 --> 00:01:39,460
there might be four changes to this genetic code table.

20
00:01:39,700 --> 00:01:44,580
So this table is pretty static, and static is pretty boring, at least for biologists.

21
00:01:44,980 --> 00:01:50,200
So we're going to ignore that, and we're actually going to focus on what we call the second genetic code.

22
00:01:50,200 --> 00:01:51,140
All right?

23
00:01:52,500 --> 00:01:57,400
And the second genetic code is essentially using tRNAs.

24
00:01:57,580 --> 00:02:00,020
So tRNAs are just RNA molecules.

25
00:02:00,100 --> 00:02:00,640
They're genes.

26
00:02:00,860 --> 00:02:04,320
They are transcribed in the same exact way as any other gene,

27
00:02:04,420 --> 00:02:05,940
but they don't get translated into proteins.

28
00:02:06,060 --> 00:02:06,720
They stay as RNA.

29
00:02:07,240 --> 00:02:15,080
And these tRNAs are essentially the link between the genome and proteins.

30
00:02:15,460 --> 00:02:19,940
So the tRNAs have this tertiary 3D structure that we see right here.

31
00:02:20,200 --> 00:02:23,840
All tRNAs have this structure, and they have this secondary structure, which is this base pairing.

32
00:02:24,200 --> 00:02:28,980
So the tRNA recognizes the mRNA with this codon sequence here,

33
00:02:29,120 --> 00:02:31,540
and they actually have an amino acid attached right here.

34
00:02:31,780 --> 00:02:32,400
All right?

35
00:02:32,440 --> 00:02:34,300
And this is going to become important later on.

36
00:02:35,140 --> 00:02:38,100
And for anyone who is not that great at protein chemistry,

37
00:02:38,320 --> 00:02:41,340
we have about 20 amino acids that make up all our proteins.

38
00:02:42,960 --> 00:02:46,840
So the big question, again, with the genetic code, second genetic code,

39
00:02:46,920 --> 00:02:50,060
is how do tRNAs get charged with certain amino acids?

40
00:02:50,200 --> 00:02:50,740
Right?

41
00:02:51,540 --> 00:02:56,040
So these tRNAs are encoded, and they don't actually carry an amino acid at first.

42
00:02:56,520 --> 00:02:58,680
So in red is the actual tRNA,

43
00:02:58,940 --> 00:03:03,140
and then the other color is the protein that charges the tRNA with the correct amino acid.

44
00:03:04,460 --> 00:03:08,160
So if we get charged with the incorrect amino acid, things go wrong.

45
00:03:08,220 --> 00:03:08,820
You probably die.

46
00:03:09,100 --> 00:03:09,520
Right?

47
00:03:09,580 --> 00:03:15,940
So tRNAs contain a set of positive features, which help the right interaction.

48
00:03:16,200 --> 00:03:17,980
So they get charged with the right amino acid.

49
00:03:18,160 --> 00:03:20,180
And they also have a set of negative correlations.

50
00:03:20,200 --> 00:03:24,520
Or negative identity features, which help them not get charged with the inappropriate.

51
00:03:25,400 --> 00:03:30,160
So what we end up with is what we've termed this tRNA interaction network.

52
00:03:30,580 --> 00:03:35,180
Since the tRNAs are confined in structure, because they have to fit inside that ribosome,

53
00:03:35,320 --> 00:03:36,440
they can't diverge.

54
00:03:36,480 --> 00:03:39,200
And the easiest thing in this cell to do, if you don't want to get charged wrong

55
00:03:39,200 --> 00:03:41,480
or have the wrong interaction, you just change shape.

56
00:03:41,820 --> 00:03:43,380
Well, tRNAs can't change shape.

57
00:03:43,420 --> 00:03:44,620
Otherwise, they don't work in translation.

58
00:03:45,100 --> 00:03:48,460
So we have what we call this, essentially, this drive to conformity,

59
00:03:48,480 --> 00:03:50,100
to maintain this shape right here.

60
00:03:50,640 --> 00:03:53,060
But at the same time, they have to be charged with the correct amino acid.

61
00:03:53,180 --> 00:03:54,400
And there's 20 different amino acids.

62
00:03:54,440 --> 00:03:58,000
So they have to have some kind of identity feature or divergence from each other.

63
00:03:58,420 --> 00:04:01,700
And this is that interaction network that dictates the second generative code.

64
00:04:01,980 --> 00:04:05,480
What features allow tRNAs to get charged from the correct amino acid?

65
00:04:06,960 --> 00:04:11,740
So when we're talking about the certain features, what we have here is the secondary structure of a tRNA.

66
00:04:11,880 --> 00:04:15,660
Which maybe if you've taken college intro biology, you've seen something like this.

67
00:04:15,720 --> 00:04:18,360
But this is essentially just a base pairing of a tRNA.

68
00:04:19,060 --> 00:04:20,180
And features we're talking about are, these are called tRNAs,

69
00:04:20,200 --> 00:04:24,240
talking about are having an A, T, C, or U at certain structural positions within the

70
00:04:24,240 --> 00:04:29,960
tRNA. So having an A at position one, or G at position one, give us a little clue about

71
00:04:29,960 --> 00:04:34,240
the function of that tRNA, actually, what amino acid it carries. So an example here

72
00:04:34,240 --> 00:04:39,100
is having, let's say, having an A at this position right here, gives us this kind of

73
00:04:39,100 --> 00:04:43,580
little bit of a probability distribution. And here we can see pretty clearly that most

74
00:04:43,580 --> 00:04:48,860
of the probability goes into an amino acid that we labeled as D. So if you have an A

75
00:04:48,860 --> 00:04:53,120
at this position, you most likely carry the amino acid D. And that's a pretty clear

76
00:04:53,120 --> 00:04:57,520
pattern, but at the same time, we have more convoluted patterns, where having an A at

77
00:04:57,520 --> 00:05:02,440
this position, again, most of the probability is in having amino acid A, but there's, you

78
00:05:02,440 --> 00:05:07,620
know, a non-zero probability of having other amino acids charged there. So the view is

79
00:05:07,620 --> 00:05:11,380
a little more convoluted and hard to understand. And then when we incorporate the fact that

80
00:05:11,380 --> 00:05:16,440
you can have different nucleotides in different positions that give us different information

81
00:05:16,440 --> 00:05:18,840
about what amino acid you carry, again, having...

82
00:05:18,860 --> 00:05:23,180
Having a G at the same position here actually gives us information that you're gonna probably

83
00:05:23,180 --> 00:05:28,200
carry an E amino acid. And over here, we get another convoluted probability distribution,

84
00:05:28,480 --> 00:05:32,320
which is a little harder to interpret. And this is gonna be difficult for us just to

85
00:05:32,320 --> 00:05:38,160
look at, because there's about 73 nucleotides in a tRNA with four possible states. So we

86
00:05:38,160 --> 00:05:43,100
have four to 73 possible combinations, which is not something we can actually interpret

87
00:05:43,100 --> 00:05:48,840
very easily, right? So how do we go about making sense of this? Well, in the back of

88
00:05:48,840 --> 00:05:53,540
the 80s, we had molecular information theory, which helped us deal with this, and it would

89
00:05:53,540 --> 00:05:58,480
strive from regular information theory. So anyone that deals with encoding or cryptography

90
00:05:59,000 --> 00:06:04,300
is probably pretty familiar with information theory, right? Proposed by Shannon in 1948,

91
00:06:04,600 --> 00:06:09,660
he defined a measure which he called entropy, right? And the actual equation for entropy

92
00:06:10,300 --> 00:06:14,120
is this lovely equation, which most people don't enjoy looking at to begin with. So

93
00:06:14,120 --> 00:06:18,740
the best way to think about entropy and actually using entropy is going through a simulation.

94
00:06:18,740 --> 00:06:18,820
So the best way to think about entropy and actually using entropy is going through a

95
00:06:18,840 --> 00:06:22,840
simulation, all right? So we run through a simulation, one that, you know, we've seen

96
00:06:22,840 --> 00:06:28,880
in programming classes all the time, teaching random number generators. We roll a fair

97
00:06:28,880 --> 00:06:32,760
six-sided die 24 times, and we're gonna pretend that we don't know how to calculate the

98
00:06:32,760 --> 00:06:37,740
probability of a die. We wanna do a simulation to figure it out, right? So we roll that die

99
00:06:38,460 --> 00:06:42,660
24 times, and we get something that may look like this, right? And the next thing we do is

100
00:06:42,660 --> 00:06:47,180
count up the number of times each number occurs. So we do that very quickly. We get,

101
00:06:47,240 --> 00:06:48,820
essentially, four. We count them up, and we get four. And we get four. And we get four. And

102
00:06:48,820 --> 00:06:49,360
we get four. And we get four. And we get four. What else do we do? Well, we do a small

103
00:06:49,360 --> 00:06:53,380
simulation, right? We have almost equal probability, represented by this bar here. But since

104
00:06:53,380 --> 00:06:58,040
it's a small simulation it's not perfect, right? Reality, it's one sixth for each side, if

105
00:06:58,040 --> 00:07:02,460
it's a fair die. If we could just run this simulation to infinity we get that proper

106
00:07:02,460 --> 00:07:08,140
probability, right? The question is how do we describe that uncertainty we have in the outcome

107
00:07:08,140 --> 00:07:13,240
of this dice roll, right? And entropy allows us to do something like that. So we have our

108
00:07:13,240 --> 00:07:18,360
fair distribution here, and we actually go through the entropy calculation. Which I've done

109
00:07:18,360 --> 00:07:20,040
done for you so no one has to do logs in their head.

110
00:07:20,800 --> 00:07:23,460
So we run this real fast and we get what's called

111
00:07:23,460 --> 00:07:25,320
2.58 bits is our entropy.

112
00:07:25,560 --> 00:07:27,700
And if you think about it, if we need six unique

113
00:07:27,700 --> 00:07:30,360
combinations, we need three bits to encode that, right?

114
00:07:30,720 --> 00:07:32,680
So we can think about bringing it right back to computer

115
00:07:32,680 --> 00:07:34,920
science and coding the same ideas, right?

116
00:07:35,480 --> 00:07:38,060
So that allows us, what kind of uncertainty do we have

117
00:07:38,060 --> 00:07:39,460
in this outcome of this die roll?

118
00:07:39,700 --> 00:07:41,580
Is measured as 2.58 bits.

119
00:07:42,280 --> 00:07:45,500
So how does this quantity change when we change

120
00:07:45,500 --> 00:07:46,260
the uncertainty?

121
00:07:46,260 --> 00:07:49,140
So if we roll a biased die where 40% of the probability

122
00:07:49,140 --> 00:07:52,140
is actually sitting inside and rolling a one,

123
00:07:52,340 --> 00:07:53,820
what does that do to our entropy calculation?

124
00:07:54,720 --> 00:07:56,960
Most cryptographers, anyone familiar,

125
00:07:57,400 --> 00:07:58,740
entropy is gonna go down, right?

126
00:07:59,080 --> 00:08:01,000
So we run our calculation again real fast.

127
00:08:01,260 --> 00:08:05,140
We go down to 2.36 bits and the whole number is 2.5 bits.

128
00:08:05,240 --> 00:08:10,700
So rolling a biased die, we're a little less shocked

129
00:08:10,700 --> 00:08:11,580
by the outcome, right?

130
00:08:11,940 --> 00:08:13,680
So we're a little less uncertain about the outcome.

131
00:08:15,000 --> 00:08:15,880
Big question is,

132
00:08:15,880 --> 00:08:17,660
how do we actually apply this to biology

133
00:08:18,240 --> 00:08:20,460
and extract information from genetic sequences?

134
00:08:21,840 --> 00:08:24,740
So we do that with what we call molecular information theory.

135
00:08:25,540 --> 00:08:27,860
We assume a lot of times that the background distribution

136
00:08:27,860 --> 00:08:29,660
of nucleotides in the genome are even.

137
00:08:29,940 --> 00:08:32,920
So there's an even amount of T's, A's, C's, and G's,

138
00:08:33,400 --> 00:08:35,680
which gives us our, what we call maximum entropy.

139
00:08:35,900 --> 00:08:37,680
We calculate that, it's two bits.

140
00:08:38,300 --> 00:08:40,300
Again, if we think about that, if we need to encode

141
00:08:40,300 --> 00:08:43,300
four unique states, we need two bits to do that, right?

142
00:08:43,640 --> 00:08:45,620
So it's running back to the same encoding theory.

143
00:08:46,280 --> 00:08:49,620
So how do we apply this idea, besides just calculating it?

144
00:08:50,180 --> 00:08:53,420
Is what we usually do is take portions of the genome,

145
00:08:53,460 --> 00:08:55,740
let's say a gene or a promoter that we're interested in,

146
00:08:55,880 --> 00:08:58,140
and we ask how variable are the positions,

147
00:08:58,500 --> 00:09:00,880
or how informative are the positions in that gene?

148
00:09:01,400 --> 00:09:05,320
So we have an example here where I've made a fake alignment

149
00:09:05,700 --> 00:09:08,200
of maybe four organisms for a promoter.

150
00:09:08,500 --> 00:09:11,060
And we actually start calculating the entropy per position.

151
00:09:11,360 --> 00:09:13,160
So down here are the probability distributions

152
00:09:13,160 --> 00:09:14,280
that we actually observed.

153
00:09:14,280 --> 00:09:16,520
And we go step-by-step through this.

154
00:09:16,600 --> 00:09:18,900
So the H here is just our maximum entropy,

155
00:09:19,340 --> 00:09:21,300
assuming the background is equal frequency,

156
00:09:22,100 --> 00:09:25,160
minus the entropy of this actual position in the alignment.

157
00:09:25,500 --> 00:09:28,200
And that is what molecular information theory people

158
00:09:28,200 --> 00:09:29,720
term information, actually.

159
00:09:30,480 --> 00:09:32,120
So we go through our calculations.

160
00:09:32,160 --> 00:09:35,020
So the first position, we have equal distribution again.

161
00:09:35,080 --> 00:09:37,280
So we have, again, maximum entropy, two bits.

162
00:09:37,440 --> 00:09:39,120
So we have two bits for maximum,

163
00:09:39,180 --> 00:09:41,440
minus two bits, equals zero bits.

164
00:09:41,680 --> 00:09:42,980
And we form these little figures,

165
00:09:43,080 --> 00:09:44,040
which we call sequence logos.

166
00:09:44,280 --> 00:09:47,300
Where at position one, we have zero bits of information,

167
00:09:47,440 --> 00:09:48,220
so we have nothing appearing.

168
00:09:48,940 --> 00:09:52,060
We essentially go and do this again for each position

169
00:09:52,700 --> 00:09:54,000
of the alignment that we have.

170
00:09:54,260 --> 00:09:56,180
So we go here, we have all As.

171
00:09:56,520 --> 00:09:58,460
So we're not uncertain about that position at all.

172
00:09:58,560 --> 00:10:00,460
So we have zero entropy in that position, actually.

173
00:10:01,320 --> 00:10:04,780
And we go two minus zero equals two bits.

174
00:10:05,040 --> 00:10:06,840
So our figure reflects that.

175
00:10:07,440 --> 00:10:10,000
We have position two going to two bits,

176
00:10:10,020 --> 00:10:13,080
and because it's only an A, we push an A.

177
00:10:13,180 --> 00:10:14,100
And we can go through this calculation,

178
00:10:14,280 --> 00:10:16,080
through all of it, but that's something

179
00:10:16,080 --> 00:10:17,860
most people can probably do at home at this point.

180
00:10:18,500 --> 00:10:21,040
So how do we actually use information theory

181
00:10:21,040 --> 00:10:22,140
to hack the generative code?

182
00:10:22,840 --> 00:10:25,280
So just a quick review.

183
00:10:26,160 --> 00:10:28,140
I remember, so we have these features,

184
00:10:28,180 --> 00:10:31,680
or states, around the tRNA that dictate function.

185
00:10:32,140 --> 00:10:33,760
And these are probability distributions.

186
00:10:34,380 --> 00:10:36,240
Now we know how to deal with probability distributions

187
00:10:36,240 --> 00:10:38,820
and calculating uncertainty based on certain features,

188
00:10:38,960 --> 00:10:39,740
or states, right?

189
00:10:39,900 --> 00:10:41,100
Before we did it with positions.

190
00:10:41,800 --> 00:10:43,320
Here we're doing it with,

191
00:10:44,280 --> 00:10:45,300
sequence alignments of tRNAs.

192
00:10:45,320 --> 00:10:49,400
So what we do is align tRNAs all together here,

193
00:10:49,920 --> 00:10:52,960
and we label the tRNAs with what amino acid they carry.

194
00:10:53,820 --> 00:10:56,760
So these carry alanine, these carry valine,

195
00:10:56,900 --> 00:10:58,540
it doesn't really matter that much,

196
00:10:58,540 --> 00:10:59,520
except that we label them correctly.

197
00:10:59,800 --> 00:11:01,200
We produce the alignments.

198
00:11:01,280 --> 00:11:02,820
And we're doing the same exact calculation.

199
00:11:03,220 --> 00:11:05,400
So maximum entropy, which in this case,

200
00:11:05,400 --> 00:11:07,260
we have 20 states, so the number of amino acids,

201
00:11:07,880 --> 00:11:10,860
minus the entropy for a certain state and position.

202
00:11:11,340 --> 00:11:13,780
So if we go through the calculation really quickly,

203
00:11:14,340 --> 00:11:15,800
if we look at position one,

204
00:11:16,620 --> 00:11:18,200
and position one having an A,

205
00:11:18,380 --> 00:11:20,820
we realize that if you have an A in position one,

206
00:11:20,860 --> 00:11:21,700
you're only in alanine.

207
00:11:21,880 --> 00:11:23,640
So here, that's about four bits of information,

208
00:11:23,880 --> 00:11:25,060
which is reflected in this graph.

209
00:11:25,440 --> 00:11:27,280
So we're not concerned at all if we know

210
00:11:27,280 --> 00:11:28,660
that you have an A in position one,

211
00:11:28,700 --> 00:11:29,280
you're in alanine.

212
00:11:29,900 --> 00:11:31,180
Same game with the C,

213
00:11:31,300 --> 00:11:32,740
if you have a C in position one,

214
00:11:33,440 --> 00:11:35,700
only valine tRNAs have a C in position one,

215
00:11:35,860 --> 00:11:38,180
so we essentially have maximum information

216
00:11:38,180 --> 00:11:39,940
at position one if there's a C.

217
00:11:40,400 --> 00:11:42,420
So we essentially start to be able to decode

218
00:11:42,780 --> 00:11:43,680
the second gender code.

219
00:11:44,280 --> 00:11:46,240
The features dictate what amino acids you get charged with.

220
00:11:47,840 --> 00:11:50,380
So eventually, when we actually run the program

221
00:11:50,380 --> 00:11:52,060
that I've created, you get these maps,

222
00:11:52,180 --> 00:11:55,280
which look confusing, but are a little better

223
00:11:55,280 --> 00:11:58,020
than just looking at four to the 73 possible combinations,

224
00:11:58,380 --> 00:11:59,380
where most of them don't matter.

225
00:11:59,660 --> 00:12:02,100
And we produce one for A, C, G, and U,

226
00:12:02,200 --> 00:12:03,000
which is working with RNA.

227
00:12:03,300 --> 00:12:05,980
And this essentially produces our structure function map.

228
00:12:06,280 --> 00:12:08,720
So your structural position one having an A,

229
00:12:08,820 --> 00:12:10,200
we know what function that gives us,

230
00:12:10,220 --> 00:12:13,480
or how confident we are about what function that gives us.

231
00:12:14,280 --> 00:12:17,160
So we can actually use these function maps

232
00:12:17,160 --> 00:12:18,100
to do applications.

233
00:12:18,780 --> 00:12:21,520
So before moving on, we'll go over the software real quick,

234
00:12:21,600 --> 00:12:23,060
if anyone wants to actually do this at home.

235
00:12:23,340 --> 00:12:25,840
So the first step you have to do in the genome

236
00:12:25,840 --> 00:12:27,140
is identify tRNA genes,

237
00:12:27,260 --> 00:12:29,820
which we use a program called tRNA-SCAN-SE,

238
00:12:30,040 --> 00:12:31,980
which will mine the genome for tRNA genes.

239
00:12:32,560 --> 00:12:34,080
Then we do a secondary alignment,

240
00:12:34,200 --> 00:12:36,160
essentially mapping to that cloverleaf shape

241
00:12:36,160 --> 00:12:37,140
that I showed previously.

242
00:12:37,380 --> 00:12:38,900
And we use a program called COVE,

243
00:12:38,940 --> 00:12:41,340
and if anyone's done any kind of natural language processing,

244
00:12:41,680 --> 00:12:43,800
it uses a context-free grammar model.

245
00:12:44,280 --> 00:12:44,760
To do that.

246
00:12:46,540 --> 00:12:48,940
And then for the information statistics and the graphics,

247
00:12:49,260 --> 00:12:52,340
we use a program which I've termed named BP Logo Fund,

248
00:12:52,460 --> 00:12:53,540
but it needs a brand new name,

249
00:12:53,620 --> 00:12:55,060
and hopefully a recursive acronym,

250
00:12:55,380 --> 00:12:56,840
so I can stay with the GNU idea.

251
00:12:57,800 --> 00:12:59,240
This is not peer-reviewed yet,

252
00:12:59,300 --> 00:13:00,420
but it's available on my GitHub,

253
00:13:00,580 --> 00:13:02,360
and will probably be published in a scientific journal

254
00:13:02,360 --> 00:13:03,240
within the end of the year.

255
00:13:03,840 --> 00:13:06,840
But we know that it's, we're pretty sure it's correct.

256
00:13:07,260 --> 00:13:10,520
So now we actually move on to the potential

257
00:13:10,520 --> 00:13:12,320
of hacking the second-gen code.

258
00:13:12,640 --> 00:13:13,560
Why do this?

259
00:13:14,280 --> 00:13:15,460
Figuring out how things work, right?

260
00:13:16,140 --> 00:13:19,440
So the big one that we actually have information from

261
00:13:19,440 --> 00:13:21,760
and data showing that it can happen and works,

262
00:13:21,920 --> 00:13:23,780
which we've developed in the lab I'm working in,

263
00:13:23,860 --> 00:13:26,160
is combating parasite infections.

264
00:13:26,660 --> 00:13:28,260
So approximately 20 million people

265
00:13:28,260 --> 00:13:30,300
have a parasite infection across the world

266
00:13:30,300 --> 00:13:32,180
at any given time, which is a lot of people.

267
00:13:33,120 --> 00:13:35,680
More than a million die annually from parasite infections,

268
00:13:36,060 --> 00:13:37,220
which is terrible.

269
00:13:37,900 --> 00:13:39,920
And this mainly happens in poor nations

270
00:13:39,920 --> 00:13:42,280
that don't have public health districts

271
00:13:42,280 --> 00:13:43,640
to clean up water

272
00:13:43,640 --> 00:13:46,540
and food sources, so Africa, Asia, Latin America.

273
00:13:46,880 --> 00:13:49,120
They already have limited access to medical care,

274
00:13:49,340 --> 00:13:52,180
and they also have a lot of parasites.

275
00:13:52,620 --> 00:13:55,300
The big issue is developing effective treatments

276
00:13:55,300 --> 00:13:56,380
for parasites have been difficult.

277
00:13:56,640 --> 00:13:58,940
The previous talk with the cancer that was up here,

278
00:13:59,100 --> 00:14:01,100
where it's essentially trying to kill the cancer

279
00:14:01,100 --> 00:14:01,920
before it kills you,

280
00:14:02,060 --> 00:14:04,640
are pretty similar to the same treatments for parasites.

281
00:14:05,140 --> 00:14:07,680
They try to kill the parasite with the drugs

282
00:14:07,680 --> 00:14:08,440
before it kills you,

283
00:14:09,080 --> 00:14:11,400
and they haven't been able to develop any new ones.

284
00:14:11,540 --> 00:14:13,160
And it's different than antibiotics

285
00:14:13,160 --> 00:14:15,460
or antibiotics attack a certain feature in bacteria

286
00:14:15,460 --> 00:14:16,280
that we don't have.

287
00:14:16,820 --> 00:14:18,720
Parasites have really similar biology,

288
00:14:18,940 --> 00:14:21,540
and there's not many target or surface area

289
00:14:21,540 --> 00:14:24,240
to attack a parasite that doesn't harm us.

290
00:14:24,840 --> 00:14:26,140
So we wanted to ask,

291
00:14:26,260 --> 00:14:28,840
do these tRNA action networks look different

292
00:14:28,840 --> 00:14:30,320
between the parasite and the human?

293
00:14:30,540 --> 00:14:32,200
And this graph looks a little complicated,

294
00:14:32,360 --> 00:14:33,980
and we used this to apply for a grant,

295
00:14:34,120 --> 00:14:35,440
but any of these colored circles

296
00:14:35,440 --> 00:14:38,340
are areas where the tRNA interaction network

297
00:14:38,340 --> 00:14:40,260
in the parasite significantly differs

298
00:14:40,260 --> 00:14:42,840
from the tRNA interaction in the network in the human.

299
00:14:43,160 --> 00:14:45,240
And these are possible areas to attack the parasite,

300
00:14:45,460 --> 00:14:46,660
while not affecting the human.

301
00:14:47,120 --> 00:14:48,700
And you go, well, that's cool,

302
00:14:48,880 --> 00:14:52,280
but we've actually teamed up with some chemists

303
00:14:52,280 --> 00:14:53,020
in Ohio State,

304
00:14:53,560 --> 00:14:55,620
and they ran small molecule analysis,

305
00:14:55,800 --> 00:14:57,980
and we found about five potential drugs

306
00:14:58,600 --> 00:15:00,680
that affect the tRNA interaction network

307
00:15:00,680 --> 00:15:01,680
within the parasite,

308
00:15:01,880 --> 00:15:04,280
and do not interfere with the tRNA interaction network

309
00:15:04,280 --> 00:15:04,680
in humans.

310
00:15:05,240 --> 00:15:07,320
And if you interrupt that tRNA interaction network

311
00:15:07,320 --> 00:15:09,220
in the parasite, it can't create proteins properly,

312
00:15:09,500 --> 00:15:11,780
and it pretty much eventually, it dies immediately.

313
00:15:12,520 --> 00:15:13,000
So,

314
00:15:13,160 --> 00:15:15,580
this is where we ran the grant,

315
00:15:15,680 --> 00:15:17,000
it'll probably move to animal phases eventually,

316
00:15:17,740 --> 00:15:19,440
but that's a real world application

317
00:15:19,440 --> 00:15:20,840
that we actually have real data for.

318
00:15:21,800 --> 00:15:23,880
So, moving on to something that we think

319
00:15:23,880 --> 00:15:25,120
might be another potential application,

320
00:15:25,340 --> 00:15:26,580
but we don't have real data for yet,

321
00:15:26,760 --> 00:15:27,920
is invasive species.

322
00:15:28,620 --> 00:15:30,760
As a biologist, we think about this all the time.

323
00:15:31,140 --> 00:15:34,680
In the US, it costs us more than $120 billion a year.

324
00:15:34,920 --> 00:15:37,880
Our loss, mainly from insect pests and fungal pathogens

325
00:15:37,880 --> 00:15:39,160
attacking our crop species.

326
00:15:39,760 --> 00:15:42,940
And again, a lot of these insects and fungal pathogens

327
00:15:43,160 --> 00:15:45,640
are very similar to the plants and very similar to us.

328
00:15:45,860 --> 00:15:47,880
So, things that kill them can be bad for us

329
00:15:47,880 --> 00:15:48,840
or bad for the plants.

330
00:15:49,300 --> 00:15:51,780
So, we wanted to add, so we're looking at,

331
00:15:51,960 --> 00:15:54,140
if we can take a pest,

332
00:15:54,280 --> 00:15:55,680
and if it's tRNA interaction network,

333
00:15:55,760 --> 00:15:56,920
again, it's significantly different

334
00:15:56,920 --> 00:15:59,240
from closely related animals in the area

335
00:15:59,240 --> 00:16:00,180
that we don't want to kill.

336
00:16:00,700 --> 00:16:03,240
We can engineer, possibly, a pesticide that can,

337
00:16:03,360 --> 00:16:06,380
you know, target specifically a certain species or organism,

338
00:16:06,580 --> 00:16:08,520
and only kill that, and actually have no harm

339
00:16:08,520 --> 00:16:11,000
to any other closely related species or humans.

340
00:16:11,360 --> 00:16:13,080
And this map is just color-coded.

341
00:16:13,160 --> 00:16:14,500
It's related to the actual potential harm

342
00:16:14,500 --> 00:16:16,080
of invasive species across the world,

343
00:16:16,780 --> 00:16:19,300
where we have kelp, we have the United States

344
00:16:19,300 --> 00:16:20,960
with the most money lost if this happens.

345
00:16:22,340 --> 00:16:24,000
So, this is another potential application

346
00:16:24,000 --> 00:16:26,540
that we're excited about and haven't fully explored yet.

347
00:16:27,380 --> 00:16:29,120
And the last one is, essentially,

348
00:16:29,200 --> 00:16:30,140
expanding the genetic code.

349
00:16:30,560 --> 00:16:32,800
So, a lot of people have thought, you know,

350
00:16:32,860 --> 00:16:35,040
they add nucleotides and try to expand

351
00:16:35,040 --> 00:16:36,680
the genetic code, give more combinations.

352
00:16:36,940 --> 00:16:39,440
Because if you have triplet, you have 64 combinations.

353
00:16:39,900 --> 00:16:41,460
But if you add two more nucleotides,

354
00:16:41,500 --> 00:16:42,400
you're going to bump up those.

355
00:16:43,160 --> 00:16:43,500
You're going to bump up those combinations, right?

356
00:16:43,620 --> 00:16:44,900
And increase coding capacity.

357
00:16:45,740 --> 00:16:47,640
But to take advantage of that coding capacity,

358
00:16:47,920 --> 00:16:50,800
you essentially have to create new tRNAs

359
00:16:50,800 --> 00:16:53,360
that get charged with those exotic amino acids,

360
00:16:53,970 --> 00:16:57,640
or it can recognize the new nucleotides

361
00:16:57,640 --> 00:16:58,680
that have been inserted into the genome.

362
00:16:58,900 --> 00:17:00,780
And this process is actually really slow.

363
00:17:01,040 --> 00:17:03,560
So, they take a tRNA and a protein that charges it

364
00:17:03,950 --> 00:17:06,360
from outside, from a different organism,

365
00:17:06,520 --> 00:17:07,860
and essentially randomly mutate it.

366
00:17:08,280 --> 00:17:11,120
Under the random mutation, they reinsert it back

367
00:17:11,120 --> 00:17:12,820
into the host cell and see if it interacts

368
00:17:13,160 --> 00:17:14,600
with that tRNA interaction network,

369
00:17:14,800 --> 00:17:16,640
and then if it can still charge the amino acid properly.

370
00:17:17,340 --> 00:17:19,320
So, creating these things can take a year,

371
00:17:19,340 --> 00:17:21,580
two years, and a couple million dollars

372
00:17:21,580 --> 00:17:23,600
worth of equipment and resources.

373
00:17:24,240 --> 00:17:26,100
But with our interaction network,

374
00:17:26,200 --> 00:17:29,340
we can actually view where a tRNA interaction network

375
00:17:29,340 --> 00:17:30,980
within a human or whatever cell we want,

376
00:17:31,180 --> 00:17:32,800
those areas are not being used,

377
00:17:33,280 --> 00:17:35,300
and can actually be exploited and mutated

378
00:17:35,300 --> 00:17:36,800
to quicken the process.

379
00:17:37,000 --> 00:17:38,420
So, we can actually direct it instead of doing

380
00:17:38,420 --> 00:17:39,220
random mutations.

381
00:17:40,120 --> 00:17:42,600
And it essentially drives down the cost pretty fast

382
00:17:42,600 --> 00:17:43,160
and the time.

383
00:17:45,120 --> 00:17:47,500
So, based on that, talk's over.

384
00:17:47,720 --> 00:17:50,760
Here's my Twitter, email, GitHub,

385
00:17:51,000 --> 00:17:51,880
if you want any of the code.

386
00:17:52,140 --> 00:17:55,740
I'll be pushing the presentation up to the GitHub

387
00:17:55,740 --> 00:17:56,760
and to my web page.

388
00:17:57,400 --> 00:17:58,700
So, there's a lot of information.

389
00:17:58,900 --> 00:17:59,960
If anyone wants to get a hold of it,

390
00:17:59,960 --> 00:18:00,760
I can actually go through it slower.

391
00:18:01,820 --> 00:18:02,960
I'm willing to talk to anybody.

392
00:18:03,180 --> 00:18:04,640
I'm an academic, so that's my job,

393
00:18:04,720 --> 00:18:05,860
is to help people.

394
00:18:07,140 --> 00:18:09,380
And I want to thank everyone for showing up

395
00:18:09,380 --> 00:18:12,400
to a talk that's relatively early for DEF CON.

396
00:18:12,400 --> 00:18:13,940
With all the drinking every night.

397
00:18:14,100 --> 00:18:16,740
And thank the biohacking team for putting this together.

398
00:18:17,360 --> 00:18:18,440
So, questions?