Responsible Genomic Data Sharing: Challenges and Approaches

Cover
Responsible Genomic Data Sharing: Challenges and Approaches
Copyright
Contributors
Section I: Privacy challenges in genomic data sharing
1. Criticality of data sharing in genomic research and public views of genomic data sharing
	1. Introduction
	2. Advancing research and scientific knowledge
	3. Importance of genomic data sharing in curing diseases
		3.1 Rare disease perspective
		3.2 Cancer perspective
		3.3 Genome-wide association studies perspective
	4. Impact of large-scale data sharing on basic research and discovery
		4.1 The International HapMap Project
		4.2 1000 Genomes Project
		4.3 The Cancer Genome Atlas project
		4.4 Encyclopedia of DNA Elements Project
		4.5 Genotype-Tissue Expression Project
	5. Reproducibility
	6. Patient/public perspective
	References
2. Genomic data access policy models
	1. Data-sharing policy developments
	2. Open-access policy model
	3. Controlled-access policy model
	4. Registered-access policy model
	5. Ongoing concerns and developments
		5.1 Maintaining consent: Consent Codes
		5.2 Data-sharing risk assessment: choosing the right access level
	References
3. Information leaks in genomic data: inference attacks
	1. Inference attacks on statistical genomic databases
	2. Inference attacks on genomic data-sharing beacons
	3. Inference attacks on kin genomic privacy
	4. Inference attacks using genotype–phenotype associations
	5. Conclusions
	References
4. Genealogical search using whole-genome genotype profiles
	1. Introduction
	2. History of personal genetic data
		2.1 HapMap
		2.2 1000 genomes Project
		2.3 UK Biobank and beyond
	3. Direct-to-consumer genetic companies
		3.1 Early days
		3.2 Growth
		3.3 Current trends
		3.4 GEDMatch and others
	4. How to encode genotype information at the genome scale
		4.1 Genotype
		4.2 Haplotype
		4.3 Phasing
		4.4 File format
			4.4.1 VCF format
			4.4.2 DTC format
			4.4.3 Compressed format
				4.4.3.1 PLINK format (BED)
				4.4.3.1 PLINK format (BED)
				4.4.3.2 UK Biobank format (BGEN)
				4.4.3.2 UK Biobank format (BGEN)
				4.4.3.3 GDS format
				4.4.3.3 GDS format
				4.4.3.4 Population-genetics inspired format: PBWT (BGT) and tree sequence (tsinfer)
				4.4.3.4 Population-genetics inspired format: PBWT (BGT) and tree sequence (tsinfer)
	5. Identity-by-descent segment and familial relatedness
		5.1 Genetic distance
		5.2 What is IBD
		5.3 Cousin nomenclature
		5.4 How IBD is related to family relationships
		5.5 Expected IBD family sharing
	6. Genealogical search
		6.1 What is a genealogy search
			6.1.1 Difference between ancestry and genealogy
		6.2 Genotype-based method
			6.2.1 HIR match
		6.3 Haplotype-based method
			6.3.1 GERMLINE
			6.3.2 PBWT-based method, RaPID and PBWT-query
			6.3.3 Refined-IBD
		6.4 Benchmarking of IBD detection: runtime, power, and accuracy
	7. Practical methods
		7.1 Methods used by DTC companies
	8. Challenges and unmet needs
		8.1 Ancestry bias
		8.2 Phasing imperfection
		8.3 Benchmarking of genealogical search
	9. Privacy concerns
	10. Conclusions
	References
Section II: Privacy-preserving techniques for responsible genomic data sharing
5. Homomorphic encryption
	1. Overview
		1.1 Early ideas
		1.2 Homomorphic encryption
		1.3 Note about terminology
		1.4 Implementations
		1.5 Standardization
		1.6 Applications
	2. Homomorphic encryption
		2.1 What is encryption?
		2.2 Partially homomorphic encryption
		2.3 Mathematical background
		2.4 (Ring) Learning With Errors
			2.4.1 Learning With Errors
			2.4.2 Ring Learning With Errors
			2.4.3 Decision problems
		2.5 The Brakerski–Fan–Vercauteren scheme
			2.5.1 Plaintexts and ciphertexts
			2.5.2 Encryption and decryption
			2.5.3 Security
			2.5.4 Noise
			2.5.5 Addition
			2.5.6 Multiplication
			2.5.7 Relinearization
			2.5.8 Implementation complexity
		2.6 Computing on encrypted integers
		2.7 Batching
			2.7.1 Micro-benchmarks
		2.8 Approximate arithmetic on encrypted numbers
		2.9 Other schemes
			2.9.1 Brakerski–Gentry–Vaikuntanathan
			2.9.2 Torus FHE
			2.9.3 High-precision arithmetic on encrypted integers
	3. Applications
		3.1 Outsourced storage and computation
		3.2 Private prediction
		3.3 Private learning
		3.4 PSI and PIR
		3.5 Biomedical applications
	4. Future Outlook
		4.1 Complexity
		4.2 Usability
		4.3 Hardware
	5. Conclusions
	References
6. Secure multi-party computation
	1. A brief overview
	2. Defining security
	3. Protocols
		3.1 Oblivious transfer
		3.2 Multiplicative triples
		3.3 Generic MPC in linear rounds
		3.4 Generic MPC in constant rounds
		3.5 Pool-based cut-and-choose
	References
7. Game theory for privacy-preserving sharing of genomic data
	1. Introduction
	2. Background
	3. Membership-inference game
		3.1 The game and its solutions
		3.2 Limitations of the model
	4. Beacon service game
		4.1 The game and its solutions
		4.2 Limitations of the model
	5. Kinship game
		5.1 The game and its solutions
		5.2 Limitations of the model
	6. Re-identification game
		6.1 The game and its solutions
		6.2 Limitations of the model
	7. Discussion
	8. Conclusions
	References
8. Trusted execution environment with Intel SGX
	1. Introduction
	2. Trusted execution environment
	3. Intel Software Guard Extensions
		3.1 Hardware architecture
			3.1.1 Memory encryption engine and protected memory region
			3.1.2 Enclave page cache and SGX data structures
			3.1.3 Enclave creation and execution
			3.1.4 Enclave paging
			3.1.5 Enclave teardown
			3.1.6 SGX platform stack
			3.1.7 Remote attestation
		3.2 Software development
		3.3 Security properties
		3.4 Performance properties
	4. HW-MPC and SGX in cloud
		4.1 TEEs in the cloud
		4.2 Developing with SGX in the cloud
		4.3 Deploying SGX applications
		4.4 Open questions on SGX in the cloud
		4.5 Putting it all together
	5. Outlook
	References
Index
	Index
	A
	B
	C
	D
	E
	F
	G
	H
	I
	J
	K
	L
	M
	N
	O
	P
	Q
	R
	S
	T
	U
	V