Secure and scalable mapping of human sequencing reads on hybrid clouds
System and methods are provided for performing privacy-preserving, high-performance, and scalable DNA read mapping on hybrid clouds including a public cloud and a private cloud. The systems and methods offer strong privacy protection and have the capacity to process millions of reads and allocate most of the workload to the public cloud at a small overall cost. The systems and methods perform seeding on the public cloud using keyed hash values of individual sequencing reads' seeds and then extend matched seeds on the private cloud. The systems and methods are designed to move the workload of read mapping from the extension stage to the seeding stage, thereby ensuring that the dominant portion of the overhead is shouldered by the public cloud.
1. One or more non-transitory, computer-readable media comprising a first plurality of instructions that, when executed by a first plurality of processors of a private computing network, causes the private computing network to:
partition each of a plurality of DNA sequence reads into a plurality of (d+2) seeds, where d is an integer;
encrypt each of the plurality of seeds;
transmit encrypted data representing each of the plurality of seeds to a public computing network;
receive from the public computing network data indicating one or more positions where a reference genome matches at least two seeds of the plurality of seeds, the at least two seeds corresponding to same DNA sequence read; and
extend the at least two seeds at each of the one or more positions where the reference genome matches the at least two seeds to determine whether the DNA sequence read corresponding to the at least two seeds aligns with the reference genome at each matched position with an edit distance less than or equal to the integer d.
2. The one or more non-transitory, computer-readable media of claim 1 , wherein the first plurality of instructions, when executed by the first plurality of processors, further causes the private computing network to:
divide the reference genome into a plurality of substrings, each of the plurality of substrings and each of the plurality of seeds being of equal length;
encrypt each unique substring of the plurality of substrings; and
transmit encrypted data representing each of the unique substrings to the public computing network.
3. The one or more non-transitory, computer-readable media of claim 2 , wherein the first plurality of instructions, when executed by the first plurality of processors, further causes the private computing network to:
encrypt each of the plurality of seeds by applying a keyed encryption algorithm to each of the plurality of seeds; and
encrypting each unique substring of the plurality of substrings by applying the keyed encryption algorithm to each unique substring.
4. The one or more non-transitory, computer-readable media of claim 2 , further comprising a second plurality of instructions that, when executed by a second plurality of processors of the public computing network, causes the public computing network to:
compare the data representing each of the plurality of seeds to the data representing each of the unique substrings; and
transmit data indicating which of the plurality of encrypted seeds matches which of the encrypted substrings to the private computing network.
5. The one or more non-transitory, computer-readable media of claim 1 , wherein the first plurality of instructions, when executed by the first plurality of processors, causes the private computing network to partition each of the plurality of DNA sequence reads into (d+2) seeds that are each between ten and twenty base pairs in length.