Un viaje por las propuestas de la ciencia ficción sobre medicina personalizada, y cómo las hemos ido cumpliendo, sobre todo gracias a las nuevas tecnologías de secuenciación en las I Jornadas Ciencia y Salud organizadas por el Ateneo Mijas.

1. ¿Ciencia ficción o medicina
personalizada? La tecnología
al servicio de la salud
M. Gonzalo Claros
Profesor titular del Dpto. de Biología Molecular y Bioquímica
Plataforma Andaluza de Bioinformática
Universidad de Málaga
#TecnoMed@MGClaros AteneoConCiencia
Se añade como aportación a la imagen y a la marca UNIVERSIDAD DE MÁLAGA
un elemento secundario de identidad, la gráfica uma.es.
El objetivo es enriquecer y actualizar la imagen de la UMA. Ambas pueden ser
usadas conjuntamente en soportes digitales (nueva web, app de la UMA).
Este recurso tipográfico es creación y propiedad de la UMA y su uso ha de regirse por
las directrices que se expecifican en este manual.
La marca uma.es no sustituirá a la marca principal salvo en aquellos casos que por
natulareza del soporte no sea viable su aplicación (superficies pequeñas o de dificil
personalización).
7
D4JUN17

2. El de la UMA, devoto de Rocío y sus iniciativas
2
#TecnoMed@MGClaros
5. UMA.ES
5.2. Variaciones de la marca uma.es
VERSIÓN DOS TINTAS EN POSITIVO
https://ateneomijas.org/ateneoconciencia

3. Para que esto no vuelva a ocurrir
3
#TecnoMed@MGClaros
5. UMA.ES
5.2. Variaciones de la marca uma.es
VERSIÓN DOS TINTAS EN POSITIVO
http://magonia.com/2015/12/03/el-archivo-del-misterio-la-homeopatia/
http://www.elconfidencial.com/tecnologia/ciencia/2017-05-27/muere-
un-nino-de-7-anos-por-combatir-una-otitis-con-homeopatia_1389764/
http://elprofedefisica.es/libro_homeopatia.htm
En las pruebas científicas no hay
que creer, del mismo modo que
nadie cuestiona que 2 + 2 = 4

4. La ciencia ficción nos vaticina el futuro
4
#TecnoMed@MGClaros
5. UMA.ES
5.2. Variaciones de la marca uma.es
VERSIÓN DOS TINTAS EN POSITIVO
http://elpais.com/elpais/2017/02/28/ciencia/1488302719_304944.html
Jules Verne
Georges Méliès

5. La medicina personalizada también tiene su ficción
5
#TecnoMed@MGClaros
5. UMA.ES
5.2. Variaciones de la marca uma.es
VERSIÓN DOS TINTAS EN POSITIVO
http://www.biotekis.es/2014/10/16/medicina-personalizada-y-cine/
Mary Wollstonecraft
Godwin
2013
2009

6. Ventaja de la medicina personalizada es clara
6
#TecnoMed@MGClaros
5. UMA.ES
5.2. Variaciones de la marca uma.es
VERSIÓN DOS TINTAS EN POSITIVO
http://cancersystemsbiology.net/predicting.html
Con la medicina personalizada

7. Ventaja de la medicina personalizada es clara
6
#TecnoMed@MGClaros
5. UMA.ES
5.2. Variaciones de la marca uma.es
VERSIÓN DOS TINTAS EN POSITIVO
http://cancersystemsbiology.net/predicting.html
Con la medicina personalizada

8. ¿Medicina personalizada o privacidad?
7
#TecnoMed@MGClaros
5. UMA.ES
5.2. Variaciones de la marca uma.es
VERSIÓN DOS TINTAS EN POSITIVO
Insuficiente para la medicina personalizada
Insuficiente para proteger la privacidad
Variantes para medicina personalizada
Privacidad

9. La percepción del riesgo nos hace tomar decisiones erradas
8
#TecnoMed@MGClaros
5. UMA.ES
5.2. Variaciones de la marca uma.es
VERSIÓN DOS TINTAS EN POSITIVO
Antenas de móviles
Centrales eléctricas
Cables de alta tensión

10. El genoma humano: 1.er paso a la medicina personalizada
9
#TecnoMed@MGClaros
5. UMA.ES
5.2. Variaciones de la marca uma.es
VERSIÓN DOS TINTAS EN POSITIVO
https://unlockinglifescode.org/timeline
1996: primer
genoma eucariota
(la levadura)
1997:
GATTACA

11. Ficción biotecnológica: GATTACA (1997)
10
#TecnoMed@MGClaros
5. UMA.ES
5.2. Variaciones de la marca uma.es
VERSIÓN DOS TINTAS EN POSITIVO
¿Qué significa
GATTACA?
http://www.smithsonianmag.com/science-nature/nasa-picks-best-amp-
worst-sci-fi-movies-what-are-yours-41527422
Una de las obras
de ciencia-
ficción con una
base científica
más sólida de la
historia del cine

12. El ADN está hecho de G, A, T y C (GATTACA)
11
#TecnoMed@MGClaros
5. UMA.ES
5.2. Variaciones de la marca uma.es
VERSIÓN DOS TINTAS EN POSITIVO
GenInicio
Guanina
Adenina
Timina
Citosina

13. Lo más «ficticio» de
12
#TecnoMed@MGClaros
5. UMA.ES
5.2. Variaciones de la marca uma.es
VERSIÓN DOS TINTAS EN POSITIVO
Hoy la boca es
fuente más simple
de ADN Identificación casi instantánea por el ADN

14. Las series de televisión han aprendido
13
#TecnoMed@MGClaros
5. UMA.ES
5.2. Variaciones de la marca uma.es
VERSIÓN DOS TINTAS EN POSITIVO

15. Lo más llamativo de
14
#TecnoMed@MGClaros
5. UMA.ES
5.2. Variaciones de la marca uma.es
VERSIÓN DOS TINTAS EN POSITIVO
Hoy la boca es
fuente más simple
de ADN Identificación casi instantánea por el ADN

16. Los biochips parecían el futuro en 1997
15
#TecnoMed@MGClaros
5. UMA.ES
5.2. Variaciones de la marca uma.es
VERSIÓN DOS TINTAS EN POSITIVO
Variantes
genéticas
Análisis
forenses
Epidemiología
Citogenética
Investigación Diagnóstico
Medicina personalizada
Identificación
Permitían analizar miles de genes a la vez

17. Obsolescencia acelerada
16
#TecnoMed@MGClaros
5. UMA.ES
5.2. Variaciones de la marca uma.es
VERSIÓN DOS TINTAS EN POSITIVO
Secuenciación de Sanger
Esto se usó para el genoma humano
s. XX
Ultrasecuenciación
s. XXI

18. Enorme salto de rendimiento
17
#TecnoMed@MGClaros
5. UMA.ES
5.2. Variaciones de la marca uma.es
VERSIÓN DOS TINTAS EN POSITIVO
http://omicsomics.blogspot.com.es/2017/01/sequencing-technology-outlook-january.html
El
rendimiento
ha pasado a
ser 108
veces mayor
Pirosecuenciación

19. Más barato. Además, más rápido
18
#TecnoMed@MGClaros
5. UMA.ES
5.2. Variaciones de la marca uma.es
VERSIÓN DOS TINTAS EN POSITIVO
https://www.genome.gov/sequencingcostsdata/
Ya podemos secuenciar un genoma humano por <1300 €
Proyecto
del genoma
humano
Pirosecuenciación
(primer método de NGS)
GA-I de Solexa
(luego Illumina) HiSeq X Ten
https://www.genome.gov/sequencingcosts/
2018:
100 €
5 días

20. ¡¡ ¿100 € un genoma? !!
19
#TecnoMed@MGClaros
5. UMA.ES
5.2. Variaciones de la marca uma.es
VERSIÓN DOS TINTAS EN POSITIVO
http://scopeblog.stanford.edu/2014/01/30/coming-soon-a-genome-
test-that-costs-less-than-a-new-pair-of-shoes/

21. ¡¡ ¿100 € un genoma? !!
19
#TecnoMed@MGClaros
5. UMA.ES
5.2. Variaciones de la marca uma.es
VERSIÓN DOS TINTAS EN POSITIVO
http://scopeblog.stanford.edu/2014/01/30/coming-soon-a-genome-
test-that-costs-less-than-a-new-pair-of-shoes/
¿Te plantea problemas éticos cuando estás
dando permiso a Google, Facebook o Twitter a
investigar tu vida gratis?

22. Los de Illumina tienen la «culpa»
20
#TecnoMed@MGClaros
5. UMA.ES
5.2. Variaciones de la marca uma.es
VERSIÓN DOS TINTAS EN POSITIVO
Cada HiSeq
X Ten genera
18 TB por
reacción
Una sola máquina de estas
genera 5 TB por reacción

23. Para hacernos una idea
21
#TecnoMed@MGClaros
5. UMA.ES
5.2. Variaciones de la marca uma.es
VERSIÓN DOS TINTAS EN POSITIVO
http://bitesizebio.com/8378/how-much-information-is-stored-in-the-human-genome/
El genoma de una
persona equivale a
1 500 millones de
caracteres de texto

24. Para hacernos una idea
21
#TecnoMed@MGClaros
5. UMA.ES
5.2. Variaciones de la marca uma.es
VERSIÓN DOS TINTAS EN POSITIVO
http://bitesizebio.com/8378/how-much-information-is-stored-in-the-human-genome/
El genoma de una
persona equivale a
1 500 millones de
caracteres de texto
Este almacenamiento
también está obsoleto

25. Acumulamos datos a mansalva
22
#TecnoMed@MGClaros
5. UMA.ES
5.2. Variaciones de la marca uma.es
VERSIÓN DOS TINTAS EN POSITIVO
https://doi.org/10.1371/journal.pbio.1002195
1-17 PB/año
1-17 × 103 TB/año
1-2 TB
1-2 × 103 GB
1-2 EB/año
1-2 × 106 TB/año
1 EB/año
106 TB/año
2-40 EB/año
2-40 × 106 TB/año
55-90 €

26. Hay «bichitos» con más ADN y más genes
23
#TecnoMed@MGClaros
5. UMA.ES
5.2. Variaciones de la marca uma.es
VERSIÓN DOS TINTAS EN POSITIVO
Protistas
Plantas
Anfibios
Peces
Insectos Mamíferos
El tamaño de
nuestro genoma
¡No somos los
amos de la Tierra!

27. ¿Quienes son más «amos» que nosotros?
24
#TecnoMed@MGClaros
5. UMA.ES
5.2. Variaciones de la marca uma.es
VERSIÓN DOS TINTAS EN POSITIVO
https://s-media-cache-ak0.pinimg.com/originals/ea/
ca/e0/eacae0476d503177b2a54f1db8d44d90.jpg
Humano
3 200 Mpb
La gran mayoría son
borradores

28. ¿Quienes son más «amos» que nosotros?
24
#TecnoMed@MGClaros
5. UMA.ES
5.2. Variaciones de la marca uma.es
VERSIÓN DOS TINTAS EN POSITIVO
https://s-media-cache-ak0.pinimg.com/originals/ea/
ca/e0/eacae0476d503177b2a54f1db8d44d90.jpg
HumanoPino
22 180 Mpb
Abeto
noruego
20 000 Mpb
3 200 Mpb
La gran mayoría son
borradores
~7×
Paris
japonica
150 000 Mpb
45×
Ameba
670 000 Mpb
210×

29. ¿Usaremos ADN en lugar de discos duros?
25
#TecnoMed@MGClaros
5. UMA.ES
5.2. Variaciones de la marca uma.es
VERSIÓN DOS TINTAS EN POSITIVO
http://revistageneticamedica.com/2017/04/19/adn-almacenamiento-datos-digitales/
Researchers Store Computer Operating System and Short Movie on DNA
New Coding Strategy Maximizes Data Storage Capacity of DNA Molecules
New York, NY - March 2, 2017 -- Humanity may soon generate more data than
hard drives or magnetic tape can handle, a problem that has scientists turning to
nature’s age-old solution for information-storage — DNA.
In a new study in Science, a pair of researchers at Columbia University and the
New York Genome Center (NYGC) show that an algorithm designed for streaming
video on a cellphone can unlock DNA’s nearly full storage potential by squeezing
more information into its four base nucleotides. They demonstrate that this
technology is also extremely reliable.
DNA is an ideal storage medium because it’s ultra-compact and can last hundreds
of thousands of years if kept in a cool, dry place, as demonstrated by the recent
recovery of DNA from the bones of a 430,000-year-old human ancestor found in a
cave in Spain.
“DNA won’t degrade over time like cassette tapes and CDs, and it won’t become
obsolete—if it does, we have bigger problems,” said study coauthor Yaniv Erlich,
a computer science professor at Columbia Engineering, a member of Columbia’s
Data Science Institute, and a core member of the NYGC.
Erlich and his colleague Dina Zielinski, an associate scientist at NYGC, chose six
files to encode, or write, into DNA: a full computer operating system, an 1895
French film, “Arrival of a train at La Ciotat,” a $50 Amazon gift card, a computer
virus, a Pioneer plaque and a 1948 study by information theorist Claude Shannon.
They compressed the files into a master file, and then split the data into short
Secuenciación
215 000 TB (2,15 × 108 GB)
de información
1 g de ADN
sin error
PCRparacopiar
1,6 bits por nucleótido
In practice, decoding took ~9 min with a Py-
thon script on a single CPU of a standard laptop
(movie S1). The decoder recovered the informa-
tion with 100% accuracy after observing only
69,870 oligos out of the 72,000 in our library
(fig. S10). To further test the robustness of our
strategy, we down-sampled the raw Illumina data
trieval of information consumes an aliquot of the
material. Copying the oligo library with PCR is
possible, but this procedure introduces noise and
induces oligo dropout. To further test the robust-
ness of our strategy, we created a deep copy of
the file by propagating the sample through nine
serial PCR amplifications (Fig. 2D). The first PCR
from an average of 7.8 perfect calls for each
per million reads (pcpm) to 4.5 pcpm in the
copy. In addition, the deep copy showed m
higher skewed representation with a neg
binomial overdispersion parameter (1/siz
0.76 compared to 0.15 in the master pool. De
the lower quality, the DNA Fountain dec
10ngoutof3µg
25µl
PCR #8PCR #7PCR #6PCR #5PCR #4PCR #3 PCR #9PCR #2PCR #1 10µl1µl1µl1µl1µl1µl1µl1µl1µl
Deepcopy
2.14Mbytetar.gz:
Perfect calls per million reads [pcpm]
Frequency
0 5 10 15 20 25 30
0.0000.0010.0020.0030.004
Mean = 7.8
Overdispersion: 0.15
Frequency
0 10 20 30 40 50 60
0.0000.0020.0040.006
Mean = 4.5
Overdispersion: 0.76
Perfect calls per million reads [pcpm]
Data payloadAdapter AdapterSeed RS5’- -3’
32Byte4Byte 2Byte
16nt24nt 8nt128nt 24nt
200nt0
72,000×
Masterpool
Perfect decoding
Sequencing on
MiSeq
25µl
MiSeq sequencing
Perfect decoding
All reads
Perfect
decoding
Perfect
decoding
All reads
Downsampling
to 5 million reads
PCR(10rounds)
Copying samples
Dilution #1
0.1x
Diluting samples
Masterpool
Masterpool
0.1x 0.1x 0.1x 0.1x 0.1x
Dilution #2 Dilution #3 Dilution #4 Dilution #5 Dilution #6 Dilution #
10ng 1ng 100pg 10pg 1pg 100fg 10fg
(215Tb/g) (2Pb/g) (21Pb/g) (215Pb/g) (2Eb/g) (21Eb/g) (215Eb/
Multiplexed
sequencing
Downsampling
to 750,000 reads
Fig. 2. Experimental setting and results for storing data on DNA.
(A) Input files for encoding, size, and type. The total amount of data was
2.14 Mbytes aftercompression.(B) Structure ofthe oligos.Black labels,length
in bytes; red, length in nucleotides; RS, Reed-Solomon error-correcting code.
(C) Experimental results of the master pool. (D) Experimental procedures
of deep copying of the oligo pool. (C and D) Histograms display the fre-
quency of perfect calls per million sequenced reads (pcpm). Red, mean;
to rescue a sufficient number of oligos. Taken
together, these results inspired us to seek a cod-
ing strategy that can better utilize the information
capacity of DNA storage devices while showing
higher data-retrieval reliability.
We devised a strategy for DNA storage, called
DNA Fountain, that approaches the Shannon ca-
pacity while providing robustness against data
corruption. Our strategy harnesses fountain codes
(18, 19), which have been developed for reliable
and effective unicasting of information over chan-
nels that are subject to dropouts, such as mobile
TV (20). In our design, we carefully adapted the
power of fountain codes to overcome both oligo
dropouts and the biochemical constraints of DNA
storage. Our encoder works in three steps (Fig.
1) (12): First, it preprocesses a binary file into a
series of nonoverlapping segments of a certain
length. Next, it iterates over two computational
steps: Luby transform and screening. The Luby
transform sets the basis for fountain codes. Basi-
cally, it packages data into any desired number
of short messages, called droplets, by selecting a
random subset of segments from the file using a
special distribution (fig. S7) and adding them
bitwise together under a binary field. The drop-
let contains two pieces of information: a data
payload part that holds the result of the addi-
tion procedure and a short, fixed-length seed.
This seed corresponds to the state of the random-
number generator of the transform during the
droplet creation and allows the decoder algo-
rithm to infer the identities of the segments in
the droplet. Theoretically, it is possible to re-
verse the Luby transform using a highly efficient
algorithm by collecting any subset of droplets
as long as the accumulated size of droplets is
slightly bigger than the size of the original file.
For DNA Fountain, our algorithm applies one
round of the transform in each iteration to create
a single droplet. Next, the algorithm moves to
the droplet screening stage. This stage is not
part of the original fountain code design and
allows us to completely realize the coding poten-
tial of each nucleotide. In screening, the algorithm
translates the binary droplet to a DNA sequence
by converting {00,01,10,11} to {A,C,G,T}, respec-
tively. Then, it screens the sequence for the
desired biochemical properties of GC content
and homopolymer runs. If the sequence passes
the screen, it is considered valid and added to
the oligo design file; otherwise, the algorithm
simply trashes the droplet. Since the Luby trans-
form can create any desired number of droplets,
we keep iterating over the droplet creation and
screening steps until a sufficient number of valid
oligos are generated. In practice, we recommend
5 to 10% more oligos than input segments (12).
Searching for valid oligos scales well with the
size of the input file and is economical for var-
ious oligo lengths within and beyond current
synthesis limits (12) (table S4).
We used DNA Fountain to encode a single
compressed file of 2,146,816 bytes in a DNA oligo
pool. The input data were in the form of a tarball
that packaged several files, including a complete
graphical operating system of 1.4 Mbytes, a movie,
and other files (12) (Fig. 2A and fig. S8). We split
the input tarball into 67,088 segments of 32 bytes
and iterated over the steps of DNA Fountain to
create valid oligos. Each droplet was 38 bytes
(304 bits): 4 bytes of the random-number gen-
erator seed, 32 bytes for the data payload, and
2 bytes for an RS error-correcting code, to reject
erroneous oligos in low-coverage conditions. With
this seed length, our strategy supports encoding
files of up to 500 Mbytes (12). The DNA oligos
had a length of 304/2 = 152 nucleotides (nt)
and were screened for homopolymer runs of
≤3 nt and GC content of 45 to 55%. We instructed
DNA Fountain to generate 72,000 oligos, yielding
a redundancy of 72,000/67,088 – 1 = 7%. We se-
lected this number of oligos due to the price
structure offered by the manufacturer, allowing
us to maximize the number of oligos per dollar.
Finally, we added upstream and downstream
annealing sites for Illumina adapters, making
our final oligos 200 nt long (Fig. 2B and fig. S9).
Encoding took 2.5 min on a single central pro-
cessing unit (CPU) of a standard laptop. We
achieved an information density of 1.57 bits/nt,
only 14% from the Shannon capacity of DNA
storage and 60% more than previous studies
with a similar scale of data (Table 1).
Sequencing and decoding the oligo pool fully
recovered the entire input file with zero errors
Fig. 1. DNA Fountain encoding. (Left) Three main algorithmic steps. (Right) Example with a small file of 32 bits. For simplicity, we partitioned the file into
eight segments of 4 bits each. The seeds are represented as 2-bit numbers and are presented for display purposes only. See (12) for the full details of each
algorithmic step.
RESEARCH | REPORT
onMarch2,2017http://science.sciencemag.org/Downloadedfrom

30. ¿Usaremos ADN en lugar de discos duros?
25
#TecnoMed@MGClaros
5. UMA.ES
5.2. Variaciones de la marca uma.es
VERSIÓN DOS TINTAS EN POSITIVO
http://revistageneticamedica.com/2017/04/19/adn-almacenamiento-datos-digitales/
Researchers Store Computer Operating System and Short Movie on DNA
New Coding Strategy Maximizes Data Storage Capacity of DNA Molecules
New York, NY - March 2, 2017 -- Humanity may soon generate more data than
hard drives or magnetic tape can handle, a problem that has scientists turning to
nature’s age-old solution for information-storage — DNA.
In a new study in Science, a pair of researchers at Columbia University and the
New York Genome Center (NYGC) show that an algorithm designed for streaming
video on a cellphone can unlock DNA’s nearly full storage potential by squeezing
more information into its four base nucleotides. They demonstrate that this
technology is also extremely reliable.
DNA is an ideal storage medium because it’s ultra-compact and can last hundreds
of thousands of years if kept in a cool, dry place, as demonstrated by the recent
recovery of DNA from the bones of a 430,000-year-old human ancestor found in a
cave in Spain.
“DNA won’t degrade over time like cassette tapes and CDs, and it won’t become
obsolete—if it does, we have bigger problems,” said study coauthor Yaniv Erlich,
a computer science professor at Columbia Engineering, a member of Columbia’s
Data Science Institute, and a core member of the NYGC.
Erlich and his colleague Dina Zielinski, an associate scientist at NYGC, chose six
files to encode, or write, into DNA: a full computer operating system, an 1895
French film, “Arrival of a train at La Ciotat,” a $50 Amazon gift card, a computer
virus, a Pioneer plaque and a 1948 study by information theorist Claude Shannon.
They compressed the files into a master file, and then split the data into short
Secuenciación
215 000 TB (2,15 × 108 GB)
de información
1 g de ADN
sin error
PCRparacopiar
1,6 bits por nucleótido
In practice, decoding took ~9 min with a Py-
thon script on a single CPU of a standard laptop
(movie S1). The decoder recovered the informa-
tion with 100% accuracy after observing only
69,870 oligos out of the 72,000 in our library
(fig. S10). To further test the robustness of our
strategy, we down-sampled the raw Illumina data
trieval of information consumes an aliquot of the
material. Copying the oligo library with PCR is
possible, but this procedure introduces noise and
induces oligo dropout. To further test the robust-
ness of our strategy, we created a deep copy of
the file by propagating the sample through nine
serial PCR amplifications (Fig. 2D). The first PCR
from an average of 7.8 perfect calls for each
per million reads (pcpm) to 4.5 pcpm in the
copy. In addition, the deep copy showed m
higher skewed representation with a neg
binomial overdispersion parameter (1/siz
0.76 compared to 0.15 in the master pool. De
the lower quality, the DNA Fountain dec
10ngoutof3µg
25µl
PCR #8PCR #7PCR #6PCR #5PCR #4PCR #3 PCR #9PCR #2PCR #1 10µl1µl1µl1µl1µl1µl1µl1µl1µl
Deepcopy
2.14Mbytetar.gz:
Perfect calls per million reads [pcpm]
Frequency
0 5 10 15 20 25 30
0.0000.0010.0020.0030.004
Mean = 7.8
Overdispersion: 0.15
Frequency
0 10 20 30 40 50 60
0.0000.0020.0040.006
Mean = 4.5
Overdispersion: 0.76
Perfect calls per million reads [pcpm]
Data payloadAdapter AdapterSeed RS5’- -3’
32Byte4Byte 2Byte
16nt24nt 8nt128nt 24nt
200nt0
72,000×
Masterpool
Perfect decoding
Sequencing on
MiSeq
25µl
MiSeq sequencing
Perfect decoding
All reads
Perfect
decoding
Perfect
decoding
All reads
Downsampling
to 5 million reads
PCR(10rounds)
Copying samples
Dilution #1
0.1x
Diluting samples
Masterpool
Masterpool
0.1x 0.1x 0.1x 0.1x 0.1x
Dilution #2 Dilution #3 Dilution #4 Dilution #5 Dilution #6 Dilution #
10ng 1ng 100pg 10pg 1pg 100fg 10fg
(215Tb/g) (2Pb/g) (21Pb/g) (215Pb/g) (2Eb/g) (21Eb/g) (215Eb/
Multiplexed
sequencing
Downsampling
to 750,000 reads
Fig. 2. Experimental setting and results for storing data on DNA.
(A) Input files for encoding, size, and type. The total amount of data was
2.14 Mbytes aftercompression.(B) Structure ofthe oligos.Black labels,length
in bytes; red, length in nucleotides; RS, Reed-Solomon error-correcting code.
(C) Experimental results of the master pool. (D) Experimental procedures
of deep copying of the oligo pool. (C and D) Histograms display the fre-
quency of perfect calls per million sequenced reads (pcpm). Red, mean;
to rescue a sufficient number of oligos. Taken
together, these results inspired us to seek a cod-
ing strategy that can better utilize the information
capacity of DNA storage devices while showing
higher data-retrieval reliability.
We devised a strategy for DNA storage, called
DNA Fountain, that approaches the Shannon ca-
pacity while providing robustness against data
corruption. Our strategy harnesses fountain codes
(18, 19), which have been developed for reliable
and effective unicasting of information over chan-
nels that are subject to dropouts, such as mobile
TV (20). In our design, we carefully adapted the
power of fountain codes to overcome both oligo
dropouts and the biochemical constraints of DNA
storage. Our encoder works in three steps (Fig.
1) (12): First, it preprocesses a binary file into a
series of nonoverlapping segments of a certain
length. Next, it iterates over two computational
steps: Luby transform and screening. The Luby
transform sets the basis for fountain codes. Basi-
cally, it packages data into any desired number
of short messages, called droplets, by selecting a
random subset of segments from the file using a
special distribution (fig. S7) and adding them
bitwise together under a binary field. The drop-
let contains two pieces of information: a data
payload part that holds the result of the addi-
tion procedure and a short, fixed-length seed.
This seed corresponds to the state of the random-
number generator of the transform during the
droplet creation and allows the decoder algo-
rithm to infer the identities of the segments in
the droplet. Theoretically, it is possible to re-
verse the Luby transform using a highly efficient
algorithm by collecting any subset of droplets
as long as the accumulated size of droplets is
slightly bigger than the size of the original file.
For DNA Fountain, our algorithm applies one
round of the transform in each iteration to create
a single droplet. Next, the algorithm moves to
the droplet screening stage. This stage is not
part of the original fountain code design and
allows us to completely realize the coding poten-
tial of each nucleotide. In screening, the algorithm
translates the binary droplet to a DNA sequence
by converting {00,01,10,11} to {A,C,G,T}, respec-
tively. Then, it screens the sequence for the
desired biochemical properties of GC content
and homopolymer runs. If the sequence passes
the screen, it is considered valid and added to
the oligo design file; otherwise, the algorithm
simply trashes the droplet. Since the Luby trans-
form can create any desired number of droplets,
we keep iterating over the droplet creation and
screening steps until a sufficient number of valid
oligos are generated. In practice, we recommend
5 to 10% more oligos than input segments (12).
Searching for valid oligos scales well with the
size of the input file and is economical for var-
ious oligo lengths within and beyond current
synthesis limits (12) (table S4).
We used DNA Fountain to encode a single
compressed file of 2,146,816 bytes in a DNA oligo
pool. The input data were in the form of a tarball
that packaged several files, including a complete
graphical operating system of 1.4 Mbytes, a movie,
and other files (12) (Fig. 2A and fig. S8). We split
the input tarball into 67,088 segments of 32 bytes
and iterated over the steps of DNA Fountain to
create valid oligos. Each droplet was 38 bytes
(304 bits): 4 bytes of the random-number gen-
erator seed, 32 bytes for the data payload, and
2 bytes for an RS error-correcting code, to reject
erroneous oligos in low-coverage conditions. With
this seed length, our strategy supports encoding
files of up to 500 Mbytes (12). The DNA oligos
had a length of 304/2 = 152 nucleotides (nt)
and were screened for homopolymer runs of
≤3 nt and GC content of 45 to 55%. We instructed
DNA Fountain to generate 72,000 oligos, yielding
a redundancy of 72,000/67,088 – 1 = 7%. We se-
lected this number of oligos due to the price
structure offered by the manufacturer, allowing
us to maximize the number of oligos per dollar.
Finally, we added upstream and downstream
annealing sites for Illumina adapters, making
our final oligos 200 nt long (Fig. 2B and fig. S9).
Encoding took 2.5 min on a single central pro-
cessing unit (CPU) of a standard laptop. We
achieved an information density of 1.57 bits/nt,
only 14% from the Shannon capacity of DNA
storage and 60% more than previous studies
with a similar scale of data (Table 1).
Sequencing and decoding the oligo pool fully
recovered the entire input file with zero errors
Fig. 1. DNA Fountain encoding. (Left) Three main algorithmic steps. (Right) Example with a small file of 32 bits. For simplicity, we partitioned the file into
eight segments of 4 bits each. The seeds are represented as 2-bit numbers and are presented for display purposes only. See (12) for the full details of each
algorithmic step.
RESEARCH | REPORT
onMarch2,2017http://science.sciencemag.org/Downloadedfrom
Problema: cuesta
10 000 $
Paradoja: ¿podremos guardar información
de 1,6 nucleótidos en tan solo 1 nt?

31. Explosión genómica
26
#TecnoMed@MGClaros
5. UMA.ES
5.2. Variaciones de la marca uma.es
VERSIÓN DOS TINTAS EN POSITIVO
https://www.achieversdaily.com/personalized-medicine-are-we-there-yet/
3 000
millones de €
1300 €
15 años
15 días
De paso, se impulsa la medicina personalizada
Menos costes no explican
este enorme incremento

32. El proceso también es más fácil
27
#TecnoMed@MGClaros
5. UMA.ES
5.2. Variaciones de la marca uma.es
VERSIÓN DOS TINTAS EN POSITIVO
https://www.genome.gov/sequencingcosts/
Human Genome Sequencing
Generating a Reference
Genome Sequence
(e.g., Human Genome Project)
Generating a Person’s
Genome Sequence
(e.g., Circa ~2016)
Genomic DNA Genomic DNA
Break genome into
large fragments and
insert into clones
Order clones
Break individual
clones into
small pieces
Generate thousands
of sequence reads
and assemble
sequence of clone
Assemble sequences
of overlapping clones
to establish
reference sequence
Break genome
into small pieces
Generate millions
of sequence reads
Align sequence reads
to established
reference sequence
Deduce starting
sequence and identify
differences from
reference sequence
Reference Sequence
Reference Sequence
....TATGCGATGCGTATTTCGTAAA....
El proyecto del genoma humano Cualquier otro humano
La referencia

33. No todo son ventajas: inseguridad
28
#TecnoMed@MGClaros
5. UMA.ES
5.2. Variaciones de la marca uma.es
VERSIÓN DOS TINTAS EN POSITIVO
Absolutamente dependiente de
los ordenadores
Hoy solo se publican borradores llenos de agujeros
Lo que nos gustaría tener
Lo que realmente tenemos

34. Un borrador es mucho mejor que nada
29
#TecnoMed@MGClaros
5. UMA.ES
5.2. Variaciones de la marca uma.es
VERSIÓN DOS TINTAS EN POSITIVO
Traslocaciones
Explica los
cambios benignos
y las
enfermedades
Polimorfismo
mononucleotídico

35. Sabemos que todos nos parecemos
30
#TecnoMed@MGClaros
5. UMA.ES
5.2. Variaciones de la marca uma.es
VERSIÓN DOS TINTAS EN POSITIVO

36. Por eso experimentamos con animales
31
#TecnoMed@MGClaros
5. UMA.ES
5.2. Variaciones de la marca uma.es
VERSIÓN DOS TINTAS EN POSITIVO
Y extrapolamos los resultados a los humanos

37. En el genoma cada vez quedan menos secretos
32
#TecnoMed@MGClaros
5. UMA.ES
5.2. Variaciones de la marca uma.es
VERSIÓN DOS TINTAS EN POSITIVO
http://book.bionumbers.org/how-many-genes-are-in-a-genome/
Los genomas contienen más
ADN superfluo que útil

38. Ya conocíamos alguna relación gen-enfermedad
33
#TecnoMed@MGClaros
5. UMA.ES
5.2. Variaciones de la marca uma.es
VERSIÓN DOS TINTAS EN POSITIVO
https://es.wikipedia.org/wiki/Enfermedad_genética
Síndrome de Down
Síndrome de Huntington
Fibrosis quística
Cáncer
Hemofilia
Diabetes de tipo 1
…

39. Descubrimos más genes relacionados con enfermedades
34
#TecnoMed@MGClaros
5. UMA.ES
5.2. Variaciones de la marca uma.es
VERSIÓN DOS TINTAS EN POSITIVO
https://clinicalscientist.wordpress.com/tag/human-genome-project/
Hay millones de variaciones
La mayoría de las
enfermedades son poligénicas

40. Genoterapia: más personalizado, imposible
35
#TecnoMed@MGClaros
5. UMA.ES
5.2. Variaciones de la marca uma.es
VERSIÓN DOS TINTAS EN POSITIVO
URL
Lentiviral vector–
mediated addition of an
antisickling β-globin
gene into autologous
hematopoietic stem
cells
Anemia
falciforme

41. Genoterapia: más personalizado, imposible
35
#TecnoMed@MGClaros
5. UMA.ES
5.2. Variaciones de la marca uma.es
VERSIÓN DOS TINTAS EN POSITIVO
URL
Lentiviral vector–
mediated addition of an
antisickling β-globin
gene into autologous
hematopoietic stem
cells
Anemia
falciforme
Este individuo sano es transgénico

42. Pero nos queda mucho por conocer
36
#TecnoMed@MGClaros
5. UMA.ES
5.2. Variaciones de la marca uma.es
VERSIÓN DOS TINTAS EN POSITIVO
http://www.seminoncol.org/article/S0093-7754(13)00216-9/pdf
Adenocarcinoma
de pulmón
Carcinoma
epidermoide

43. >30% del cáncer se debe a mutaciones imprevisibles
37
#TecnoMed@MGClaros
5. UMA.ES
5.2. Variaciones de la marca uma.es
VERSIÓN DOS TINTAS EN POSITIVO
Tomasetti et al., Science 355, 1330–1334 (2017)
No lo estamos
haciendo tan mal

44. Se pueden analizar muchos genomas a la vez
38
#TecnoMed@MGClaros
5. UMA.ES
5.2. Variaciones de la marca uma.es
VERSIÓN DOS TINTAS EN POSITIVO
URL
Sequencing thousands of single-cell genomes with combinatorial indexing
Sarah A Vitak, Kristof A Torkenczy, Jimi L Rosenkrantz, Andrew J Fields, Lena
Christiansen, Melissa H Wong, Lucia Carbone, Frank J Steemers & Andrew Adey
Nature Methods 14, 302–308 (2017) doi:10.1038/nmeth.4154
Adenocarcinoma pancreático (PDAC) en estadio III
Se analizaron más de
16.500 células normales y
tumorales
Se diferenciaron cuatro poblaciones
diferentes en el mismo tumor, cada
una con sus variantes específicas

45. La capacidad predictiva va mejorando
39
#TecnoMed@MGClaros
5. UMA.ES
5.2. Variaciones de la marca uma.es
VERSIÓN DOS TINTAS EN POSITIVO
Un número
significativo de
variantes del
paciente le
predisponen a
una
enfermedad
Sus variantes
no están
asociadas a
ninguna
enfermedad ni
al riesgo de
padecerla

46. Se han aprobado unos pocos kits predictivos
40
#TecnoMed@MGClaros
5. UMA.ES
5.2. Variaciones de la marca uma.es
VERSIÓN DOS TINTAS EN POSITIVO
https://www.fda.gov/MedicalDevices/
ProductsandMedicalProcedures/InVitroDiagnostics/ucm330711.htm
Es terreno abonado para
los charlatanes
La mayoría, para el cáncer

47. 41
#TecnoMed@MGClaros
5. UMA.ES
5.2. Variaciones de la marca uma.es
VERSIÓN DOS TINTAS EN POSITIVO
http://francis.naukas.com/2014/01/06/francis-en-rosavientos-noticias-para-manana-sabado-2/

48. 41
#TecnoMed@MGClaros
5. UMA.ES
5.2. Variaciones de la marca uma.es
VERSIÓN DOS TINTAS EN POSITIVO
http://francis.naukas.com/2014/01/06/francis-en-rosavientos-noticias-para-manana-sabado-2/
Gen saltarín
Neuron 81, 306–313, January 22, 2014
http://dx.doi.org/10.1016/j.neuron.2013.10.053
Síndrome de Rett
Ataxia telangiectasia
Cáncer de próstata
LINE1 (L1) está relacionado
con más enfermedades

49. 41
#TecnoMed@MGClaros
5. UMA.ES
5.2. Variaciones de la marca uma.es
VERSIÓN DOS TINTAS EN POSITIVO
http://francis.naukas.com/2014/01/06/francis-en-rosavientos-noticias-para-manana-sabado-2/
Gen saltarín
Neuron 81, 306–313, January 22, 2014
http://dx.doi.org/10.1016/j.neuron.2013.10.053
Síndrome de Rett
Ataxia telangiectasia
Cáncer de próstata
LINE1 (L1) está relacionado
con más enfermedades
Lo que te da la vida,
también te la podría quitar

50. Nuestro genoma es una bomba de relojería
42
#TecnoMed@MGClaros
5. UMA.ES
5.2. Variaciones de la marca uma.es
VERSIÓN DOS TINTAS EN POSITIVO
¡Tenemos más
transposones
que genes!
Procura no darle al interruptor de la cuenta atrás
Genoma nuclear humano

51. Tampoco es que todo sea NUESTRO genoma
43
#TecnoMed@MGClaros
5. UMA.ES
5.2. Variaciones de la marca uma.es
VERSIÓN DOS TINTAS EN POSITIVO
https://www.snpedia.com/index.php/Obesity
Variantes relacionadas
con la obesidad
Epigenética

52. Ambiente y microbiota siguen contando
44
#TecnoMed@MGClaros
5. UMA.ES
5.2. Variaciones de la marca uma.es
VERSIÓN DOS TINTAS EN POSITIVO
Microbiota: conjunto de
microorganismos que
viven en el interior de un
ser vivo normal

53. El experimento revelador
45
#TecnoMed@MGClaros
5. UMA.ES
5.2. Variaciones de la marca uma.es
VERSIÓN DOS TINTAS EN POSITIVO
URL
Gemelo obeso
Gemelo delgado
Ratón delgado
Ratón obeso
Ratón normal
Trasplante de
microbiota
Dieta con poca grasa
y mucha fibra
Extracción de
microbiota
Ratón delgadoRatón obeso

54. La microbiota altera los tratamientos
46
#TecnoMed@MGClaros
5. UMA.ES
5.2. Variaciones de la marca uma.es
VERSIÓN DOS TINTAS EN POSITIVO
Article
Host-Microbe Co-metabolism Dictates Cancer Drug
Efficacy in C. elegans
Graphical Abstract
Highlights
d Drug-microbe-host high-throughput screens reveal new
mechanisms for cancer drugs
d Microbes integrate nutritional and drug cues regulating
treatment efficacy in the host
d Ribonucleotide co-metabolism of cancer pro-drugs exists
between host and microbe
d Imbalanced bacterial deoxynucleotides synergize 5-FU-
induced autophagic cell death
Authors
Timothy A. Scott, Leonor M. Quintaneiro,
Povilas Norvaisas, ..., Athanasios Typas,
Nicholas D.E. Greene, Filipe Cabreiro
Correspondence
f.cabreiro@ucl.ac.uk
In Brief
A three-way high-throughput screen
involving host-microbe-drug interactions
reveals that the beneficial impact of some
drugs can be due to effects of drug-
dependent alterations by gut microbe
composition rather than direct action of
the therapeutic itself.
HOLOBIONT
C. elegans
E. coli
Microbiota
Host
Drugconversion
Cancer survival
ImprovedUnaltered
DRUG
5-FU
Ribonucleotide
metabolism
Microbiota
conversionDrDugc
dRibonucleotide
metabolism
NUTRITION
Vitamin B6/B9
UMP precursors
Scott et al., 2017, Cell 169, 442–456
April 20, 2017 ª 2017 The Author(s). Published by Elsevier Inc.
http://dx.doi.org/10.1016/j.cell.2017.03.040
Scott et al., 2017, Cell 169, 442–456
Organismo modelo
Fármaco
Las diferencias en la
población de bacterias
del intestino explicaría
que algunos
quimioterápicos no
funcionen en ciertos
pacientes

55. Estamos llenos de bacterias que tienen su propio genoma
47
#TecnoMed@MGClaros
5. UMA.ES
5.2. Variaciones de la marca uma.es
VERSIÓN DOS TINTAS EN POSITIVO
URL
Por dentro… …y por fuera
Nature Reviews Microbiology 9, 244-253 (April 2011)
Nature 449, 811-818 (18 October 2007)

56. Estamos llenos de bacterias que tienen su propio genoma
47
#TecnoMed@MGClaros
5. UMA.ES
5.2. Variaciones de la marca uma.es
VERSIÓN DOS TINTAS EN POSITIVO
URL
Por dentro… …y por fuera
Nature Reviews Microbiology 9, 244-253 (April 2011)
Nature 449, 811-818 (18 October 2007)
Responsables, entre otras
muchas cosas, del olor corporal

57. Lo sabemos gracias a la metagenómica
48
#TecnoMed@MGClaros
5. UMA.ES
5.2. Variaciones de la marca uma.es
VERSIÓN DOS TINTAS EN POSITIVO
https://bacpathgenomics.wordpress.com/tag/science/

58. Lo sabemos gracias a la metagenómica
48
#TecnoMed@MGClaros
5. UMA.ES
5.2. Variaciones de la marca uma.es
VERSIÓN DOS TINTAS EN POSITIVO
https://bacpathgenomics.wordpress.com/tag/science/
Debemos abandonar la idea de enfermedades
poligénicas para convencernos que, en realidad,
las enfermedades son poligenómicas
Conocemos una cantidad ridícula
de especies de microorganismos

59. Futuro: curar los genes de «algún» genoma
49
#TecnoMed@MGClaros
5. UMA.ES
5.2. Variaciones de la marca uma.es
VERSIÓN DOS TINTAS EN POSITIVO
D4JUN17
http://about.me/mgclaros/
AteneoConCiencia