SLIDE 1
Molecular simulations of DNA loop extrusion explain and predict - - PowerPoint PPT Presentation
Molecular simulations of DNA loop extrusion explain and predict - - PowerPoint PPT Presentation
Molecular simulations of DNA loop extrusion explain and predict human genome architecture Adrian Sanborn Stanford University; The Center for Genome Architecture at Baylor College of Medicine, Rice University FUNDAMENTAL PUZZLE: SAME GENOME
SLIDE 2
SLIDE 3
THE GENOME IS FULL OF “SPOOKY ACTION AT A DISTANCE”
SLIDE 4
THE GENOME IS FULL OF “SPOOKY ACTION AT A DISTANCE”
SLIDE 5
…CGTTTACGAAAATCGCAAAACTTTCGATACCCATAGGCTACTGATCATACGACCGTTTACGAAAATCGAAACCTTTCCGATCTAGGCTAC…
3 BILLION LETTERS 2 METERS
Nucleus
Cell
6 μm
THE HUMAN GENOME IS LONG
SLIDE 6
10 bp 100 bp 1 Kb 10 Kb 100 Kb 1 Mb 10 Mb 100 Mb
SLIDE 7
10 bp 100 bp 1 Kb 10 Kb 100 Kb 1 Mb 10 Mb 100 Mb
SLIDE 8
HOW DOES THE GENOME FOLD?
- 1. Experimental Technology
- 2. Biology
- 3. Physics
SLIDE 9
PART I: EXPERIMENTAL TECHNOLOGY
SLIDE 10
MICROSCOPY & FLUORESCENT IN SITU HYBRIDIZATION
FISH
SLIDE 11
MICROSCOPY & FLUORESCENT IN SITU HYBRIDIZATION
FISH
SLIDE 12
Lieberman-Aiden, van Berkum et al. Science 2009
HI-C MEASURES SPATIAL PROXIMITY OF THE FOLDED GENOME
SLIDE 13
CONTACT MAPPING
Exploring structure via proximity
SLIDE 14
SLIDE 15
SLIDE 16
SLIDE 17
SLIDE 18
SLIDE 19
SLIDE 20
# of Pictures Together
SIMPSONS
CONTACT MAP
SLIDE 21
Lieberman-Aiden, van Berkum et al. Science 2009
HI-C MEASURES SPATIAL PROXIMITY OF THE FOLDED GENOME
SLIDE 22
Lieberman-Aiden, van Berkum et al. Science 2009
HI-C MEASURES SPATIAL PROXIMITY OF THE FOLDED GENOME
SLIDE 23
Hi-C
GENERATES GENOME- WIDE CONTACT MAPS
SLIDE 24
Chromosome
Hi-C
GENERATES GENOME- WIDE CONTACT MAPS
SLIDE 25
Hi-C
GENERATES GENOME- WIDE CONTACT MAPS
Genome
SLIDE 26
Genome
Hi-C
GENERATES GENOME- WIDE CONTACT MAPS
SLIDE 27
Genome
Hi-C
GENERATES GENOME- WIDE CONTACT MAPS
SLIDE 28
Genome
Hi-C
GENERATES GENOME- WIDE CONTACT MAPS
SLIDE 29
Hi-C
GENERATES GENOME- WIDE CONTACT MAPS
SLIDE 30
Chromosome 8
Hi-C
GENERATES GENOME- WIDE CONTACT MAPS
700 Reads/250 kb2
SLIDE 31
Hi-C
GENERATES GENOME- WIDE CONTACT MAPS
700 Reads/250 kb2
SLIDE 32
A A
Hi-C
GENERATES GENOME- WIDE CONTACT MAPS
700 Reads/250 kb2
SLIDE 33
A A
Hi-C
GENERATES GENOME- WIDE CONTACT MAPS
700 Reads/250 kb2
SLIDE 34
A B A B
Hi-C
GENERATES GENOME- WIDE CONTACT MAPS
700 Reads/250 kb2
SLIDE 35
30 million contacts
NEW CONTACT MAPS ARE AT KILOBASE RESOLUTION
SLIDE 36
5 billion contacts 30 million contacts
NEW CONTACT MAPS ARE AT KILOBASE RESOLUTION
Rao & Huntley, et al., Cell 2014
SLIDE 37
PART II: BIOLOGY
Mary Ellen Scherl
SLIDE 38
LOOPS IN CHROMATIN = PEAKS IN A CONTACT MAP
A A-2ε A-ε A+ε A+2ε B-ε B-2ε B B+ε B+2ε
SLIDE 39
LOOPS IN CHROMATIN = PEAKS IN A CONTACT MAP
A A-2ε A-ε A+ε A+2ε B-ε B-2ε B B+ε B+2ε
SLIDE 40
LOOPS IN CHROMATIN = PEAKS IN A CONTACT MAP
A A-2ε A-ε A+ε A+2ε B-ε B-2ε B B+ε B+2ε
SLIDE 41
THE GENOME HAS ~10,000 LOOPS.
Rao & Huntley, et al., Cell 2014
SLIDE 42
THE GENOME HAS ~10,000 LOOPS.
Rao & Huntley, et al., Cell 2014
SLIDE 43
THE GENOME HAS ~10,000 LOOPS.
Rao & Huntley, et al., Cell 2014
SLIDE 44
LOOPS ARE OFTEN CONSERVED ACROSS CELL TYPE…
SLIDE 45
LOOPS ARE OFTEN CONSERVED ACROSS CELL TYPE…
SLIDE 46
LOOPS ARE OFTEN CONSERVED ACROSS CELL TYPE…
Rao & Huntley, et al., Cell 2014
SLIDE 47
…AND SPECIES.
SLIDE 48
…AND SPECIES.
Rao & Huntley, et al., Cell 2014
SLIDE 49
LOOPS TURN GENES ON AND OFF .
SLIDE 50
LOOPS TURN GENES ON AND OFF .
Rao & Huntley, et al., Cell 2014
SLIDE 51
LOOPS TURN GENES ON AND OFF .
SLIDE 52
LOOPS TURN GENES ON AND OFF .
Rao & Huntley, et al., Cell 2014
SLIDE 53
LOOPS DEMARCATE CONTACT DOMAINS…
Rao & Huntley, et al., Cell 2014
SLIDE 54
LOOPS DEMARCATE CONTACT DOMAINS…
Rao & Huntley, et al., Cell 2014
SLIDE 55
…WHICH PARTITION THE GENOME
SLIDE 56
…WHICH PARTITION THE GENOME
Rao & Huntley, et al., Cell 2014
SLIDE 57
HOW DO LOOPS AND DOMAINS FORM?
SLIDE 58
PART III: PHYSICS
SLIDE 59
CONTACT PROBABILITY AS A FUNCTION OF GENOMIC DISTANCE
SLIDE 60
CONTACT PROBABILITY AS A FUNCTION OF GENOMIC DISTANCE
SLIDE 61
FOLDED DNA IS NOT AT EQUILIBRIUM
Erez Lieberman-Aiden, Nynke van Berkum et al. Science 2009
SLIDE 62
FOLDED DNA IS NOT AT EQUILIBRIUM
Erez Lieberman-Aiden, Nynke van Berkum et al. Science 2009
SLIDE 63
Sanborn & Rao, et al., PNAS, 2015
SLIDE 64
Sanborn & Rao, et al., PNAS, 2015
CONTACT PROBABILITY EXPONENT WITHIN DOMAINS IS -0.75
SLIDE 65
Sanborn & Rao, et al., PNAS, 2015
CONTACT PROBABILITY EXPONENT WITHIN DOMAINS IS -0.75
SLIDE 66
Sanborn & Rao, et al., PNAS, 2015
CONTACT PROBABILITY EXPONENT WITHIN DOMAINS IS -0.75
SLIDE 67
Sanborn & Rao, et al., PNAS, 2015
INTERNAL VERSUS EXTERNAL FORCES
SLIDE 68
Sanborn & Rao, et al., PNAS, 2015
INTERNAL FORCES à LINEAR COLLAPSE
SLIDE 69
TENSION GLOBULE RECAPITULATES OBSERVED CONTACT PROBABILITY
Sanborn & Rao, et al., PNAS, 2015
Hi-C Tension globule
SLIDE 70
INTERNAL FORCES ARE COMPUTATIONALLY INTENSIVE
SLIDE 71
INTERNAL FORCES ARE COMPUTATIONALLY INTENSIVE
Polymer backbone Lennard-Jones forces
Simulate Langevin dynamics using LAMMPS package
SLIDE 72
INTERNAL FORCES ARE COMPUTATIONALLY INTENSIVE
- Simulations up to 50,000 beads (50 megabases)
- Forces must be calculated between all pairs of beads
- Hundreds of replicate structures must be simulated
- GPU parallelization à ~10x speed-up
SLIDE 73
LOOPS AND DOMAINS FORM WITHIN TENSION GLOBULES
SLIDE 74
LOOPS ARE ANCHORED AT CONVERGENT CTCF SITES
SLIDE 75
Rao & Huntley, et al., Cell 2014
LOOPS ARE ANCHORED AT CONVERGENT CTCF SITES
SLIDE 76
LOOP FORMATION BY DIFFUSION
Sanborn & Rao, et al., PNAS, 2015
SLIDE 77
LOOP FORMATION BY DIFFUSION
Sanborn & Rao, et al., PNAS, 2015
SLIDE 78
LOOP FORMATION BY DIFFUSION
Sanborn & Rao, et al., PNAS, 2015
SLIDE 79
LOOP FORMATION BY DIFFUSION
Sanborn & Rao, et al., PNAS, 2015
SLIDE 80
LOOP FORMATION BY DIFFUSION
Sanborn & Rao, et al., PNAS, 2015
SLIDE 81
LOOP FORMATION BY DIFFUSION
Sanborn & Rao, et al., PNAS, 2015
SLIDE 82
DIFFUSION DOES NOT EXPLAIN CONVERGENT RULE
Sanborn & Rao, et al., PNAS, 2015
SLIDE 83
DIFFUSION DOES NOT EXPLAIN CONVERGENT RULE
Sanborn & Rao, et al., PNAS, 2015
SLIDE 84
DIFFUSION DOES NOT EXPLAIN WHY LOOPS TEND NOT TO OVERLAP
Sanborn & Rao, et al., PNAS, 2015
SLIDE 85
DIFFUSION DOES NOT EXPLAIN LACK OF FOCAL CONTACTS BETWEEN CHROMOSOMES
Sanborn & Rao, et al., PNAS, 2015
SLIDE 86
DIFFUSION LEADS TO ENTANGLEMENT
Sanborn & Rao, et al., PNAS, 2015
SLIDE 87
LOOPS AND DOMAINS FORM VIA EXTRUSION
Najeeb Tarazi, Adrian Sanborn
SLIDE 88
LOOPS AND DOMAINS FORM VIA EXTRUSION
Najeeb Tarazi, Adrian Sanborn
SLIDE 89
LOOPS AND DOMAINS FORM VIA EXTRUSION
Najeeb Tarazi, Adrian Sanborn
SLIDE 90
LOOPS AND DOMAINS FORM VIA EXTRUSION
Sanborn, Rao et al., PNAS 2015
SLIDE 91
LOOPS AND DOMAINS FORM VIA EXTRUSION
Sanborn, Rao et al., PNAS 2015
SLIDE 92
EXTRUSION EXPLAINS MANY KEY OBSERVATIONS
Sanborn, Rao et al., PNAS 2015
SLIDE 93
EXTRUSION EXPLAINS MANY KEY OBSERVATIONS
Sanborn, Rao et al., PNAS 2015
Loops at convergent CTCF sites
SLIDE 94
EXTRUSION EXPLAINS MANY KEY OBSERVATIONS
Sanborn, Rao et al., PNAS 2015
Loops at convergent CTCF sites Loops don’t overlap
SLIDE 95
EXTRUSION EXPLAINS MANY KEY OBSERVATIONS
Sanborn, Rao et al., PNAS 2015
Loops at convergent CTCF sites Loops don’t overlap Loops are intrachromosomal
SLIDE 96
EQUATIONS FOR SIMULATING EXTRUSION
Polymer backbone Lennard-Jones forces Loop bonds
SLIDE 97
A DAY IN THE LIFE OF AN EXTRUSION COMPLEX
Najeeb Tarazi, Adrian Sanborn
SLIDE 98
A DAY IN THE LIFE OF AN EXTRUSION COMPLEX
Najeeb Tarazi, Adrian Sanborn
SLIDE 99
A DAY IN THE LIFE OF AN EXTRUSION COMPLEX
Najeeb Tarazi, Adrian Sanborn
SLIDE 100
LOOP EXTRUSION FORMS SPATIALLY SEGREGATED DOMAINS
Najeeb Tarazi, Adrian Sanborn
SLIDE 101
LOOP EXTRUSION FORMS SPATIALLY SEGREGATED DOMAINS
Najeeb Tarazi, Adrian Sanborn
SLIDE 102
LOOP EXTRUSION FORMS SPATIALLY SEGREGATED DOMAINS
Najeeb Tarazi, Adrian Sanborn
SLIDE 103
LOOP EXTRUSION FORMS CHROMOSOME TERRITORIES
Najeeb Tarazi, Adrian Sanborn
SLIDE 104
LOOP EXTRUSION FORMS CHROMOSOME TERRITORIES
Najeeb Tarazi, Adrian Sanborn
SLIDE 105
LOOP EXTRUSION FORMS CHROMOSOME TERRITORIES
Najeeb Tarazi, Adrian Sanborn
SLIDE 106
EXTRUSION ENABLES PREDICTION OF HI-C MAPS IN SILICO FROM CTCF CHIP-SEQ DATA
Sanborn, Rao et al., PNAS 2015
SLIDE 107
EXTRUSION ENABLES PREDICTION OF HI-C MAPS IN SILICO FROM CTCF CHIP-SEQ DATA
Sanborn, Rao et al., PNAS 2015
SLIDE 108
EXTRUSION ENABLES PREDICTION OF HI-C MAPS IN SILICO FROM CTCF CHIP-SEQ DATA
Sanborn, Rao et al., PNAS 2015
SLIDE 109
EXTRUSION ENABLES PREDICTION OF HI-C MAPS IN SILICO FROM CTCF CHIP-SEQ DATA
Sanborn, Rao et al., PNAS 2015
SLIDE 110
EXTRUSION ENABLES PREDICTION OF HI-C MAPS IN SILICO FROM CTCF CHIP-SEQ DATA
Sanborn, Rao et al., PNAS 2015
SLIDE 111
EXTRUSION ENABLES PREDICTION OF HI-C MAPS IN SILICO FROM CTCF CHIP-SEQ DATA
Sanborn, Rao et al., PNAS 2015
SLIDE 112
EXTRUDED DOMAINS RECAPITULATE CONTACT PROBABILITY EXPONENT
Sanborn & Rao, et al., PNAS, 2015
Hi-C Extrusion
SLIDE 113
EXTRUSION SIMULATIONS PREDICT GENOME REFOLDING UPON EDITING OF CTCF MOTIFS
Sanborn, Rao et al., PNAS 2015
SLIDE 114
EXTRUSION SIMULATIONS PREDICT GENOME REFOLDING UPON EDITING OF CTCF MOTIFS
Sanborn, Rao et al., PNAS 2015
SLIDE 115
EXTRUSION SIMULATIONS PREDICT GENOME REFOLDING UPON EDITING OF CTCF MOTIFS
Sanborn, Rao et al., PNAS 2015
SLIDE 116
EXTRUSION SIMULATIONS PREDICT GENOME REFOLDING UPON EDITING OF CTCF MOTIFS
Sanborn, Rao et al., PNAS 2015
SLIDE 117
EXTRUSION SIMULATIONS PREDICT GENOME REFOLDING UPON EDITING OF CTCF MOTIFS
Sanborn, Rao et al., PNAS 2015
SLIDE 118
EXTRUSION SIMULATIONS PREDICT GENOME REFOLDING UPON EDITING OF CTCF MOTIFS
Sanborn, Rao et al., PNAS 2015
SLIDE 119
EXTRUSION SIMULATIONS PREDICT GENOME REFOLDING UPON EDITING OF CTCF MOTIFS
Sanborn, Rao et al., PNAS 2015
SLIDE 120
EXTRUSION SIMULATIONS PREDICT GENOME REFOLDING UPON EDITING OF CTCF MOTIFS
Sanborn, Rao et al., PNAS 2015
SLIDE 121
EXTRUSION SIMULATIONS PREDICT GENOME REFOLDING UPON EDITING OF CTCF MOTIFS
Sanborn, Rao et al., PNAS 2015
SLIDE 122
EXTRUSION SIMULATIONS PREDICT GENOME REFOLDING UPON EDITING OF CTCF MOTIFS
Sanborn, Rao et al., PNAS 2015
SLIDE 123
EXTRUSION SIMULATIONS PREDICT GENOME REFOLDING UPON EDITING OF CTCF MOTIFS
Sanborn, Rao et al., PNAS 2015
SLIDE 124
EXTRUSION SIMULATIONS PREDICT GENOME REFOLDING UPON EDITING OF CTCF MOTIFS
Sanborn, Rao et al., PNAS 2015
SLIDE 125
EXTRUSION SIMULATIONS PREDICT GENOME REFOLDING UPON EDITING OF CTCF MOTIFS
Sanborn, Rao et al., PNAS 2015
SLIDE 126
Sanborn & Rao, et al., PNAS, 2015
EXTRUSION SIMULATIONS PREDICT GENOME REFOLDING UPON EDITING OF CTCF MOTIFS
SLIDE 127
GOING FORWARD
SLIDE 128
GOING FORWARD
- Loop extrusion model shows tremendous promise
SLIDE 129
GOING FORWARD
- Loop extrusion model shows tremendous promise
- Next steps:
- Computationally generate 3D maps for many,
many new cell types
- Release software for predicting how genomes
refold upon editing
SLIDE 130
MAJOR CHALLENGE
SLIDE 131
MAJOR CHALLENGE
- Bond shifting disallowed during simulation
SLIDE 132
MAJOR CHALLENGE
- Bond shifting disallowed during simulation
- Extrusion does not parallelize well on GPUs
SLIDE 133
MAJOR CHALLENGE
- Bond shifting disallowed during simulation
- Extrusion does not parallelize well on GPUs
- Need to redesign GPU implementation to integrate
bond shifting
SLIDE 134
MAJOR CHALLENGE
- Bond shifting disallowed during simulation
- Extrusion does not parallelize well on GPUs
- Need to redesign GPU implementation to integrate
bond shifting
- Sound interesting?
SLIDE 135
Mary Ellen Scherl
People
Suhas Rao* Su-Chen Huang Neva Durand Miriam Huntley* Andrew Jewett Ivan Bochkov Dharmaraj Chinnappan Ashok Cutkosky Jian Li Kristopher Geeting Andreas Gnirke Alexandre Melnikov Doug McKenna Elena Stamenova Eric Lander Erez Lieberman Aiden