Genomic services system with dual-phase genotype imputation
Systems, platforms, methods and media for providing genomic services are disclosed. In one example, a genomic services platform comprises a network interface through which are received genomic sequence reads derived from a biological sample obtained from a user. The platform also includes a bioinformatics processing pipeline including a read alignment module configured to generate observed sequence data by aligning the sequence reads relative to a reference sequence, a variant calling module operative to identify observed variants in the observed sequence data, and a variant refinement module for producing genotype data including a set of refined variants associated with the user. A variant imputation module produces a set of imputed variants associated with the user, and is configured to receive, as input, at least some of the genotype data and separate the genotype data into high-quality and low-quality genotypes based on a genotype quality.
1. A system for providing genomic services, the system comprising:
a network interface through which genomic sequence reads of a user are received, the genomic sequence reads of the user having been generated by genomic sequencing equipment based on a biological sample provided by the user;
a set of server computers storing instructions that, when executed by one or more processors of the set of server computers causes the set of server computers to perform operations comprising:
aligning the received genomic sequence reads of the user to a reference human genome;
identifying genomic variants of the user based on a comparison of the aligned sequence reads of the user to the reference human genome;
generating a pre-imputation file that corresponds to the user and specifies genotypes of the user, the pre-imputation file specifying the genotypes of the user by specifying the identified genomic variants of the user, the pre-imputation file indicating a high-quality subset of genotypes whose error probabilities do not transgress a predetermined threshold probability, the pre-imputation file indicating a low-quality subset of genotypes whose error probabilities transgress the predetermined threshold probability;
performing a first imputation that adds first additional genotypes to the genotypes specified in the pre-imputation file based on the high-quality subset of genotypes and without influence from the low-quality subset of genotypes, the first imputation modifying at least some of the error probabilities of the low-quality subset of genotypes;
performing a second imputation that adds second additional genotypes to the genotypes specified in the pre-imputation file based on both the high-quality subset of genotypes whose error probabilities do not transgress the predetermined threshold probability and the low-quality subset of genotypes at least some of whose error probabilities were modified by the performed first imputation; and
generating a post-imputation file that corresponds to the user, specifies the genotypes specified in the pre-imputation file, and further specifies the first and second additional genotypes of the user, based on the performed first and second imputations.
2. The genomic services system of claim 1 , wherein the operations further comprise:
determining error probabilities of the genotypes specified by the pre-imputation file, the determined error probabilities including the error probabilities of the high-quality-subset of genotypes and the error probabilities of the low-quality subset of genotypes.
3. The genomic services system of claim 2 , wherein:
variations in the determined error probabilities of the genotypes specified by the pre-imputation file indicate variations in depth of coverage among the genomic sequence reads generated by the genomic sequencing equipment.
4. The genomic services system of claim 1 , wherein:
the first imputation is included in a first phase of a dual-phase imputation in which the first phase provides results to a second phase of the dual-phase imputation; and
the second imputation is included in the second phase of the dual-phase imputation and performed based on the results provided by the first phase of the dual-phase imputation.
5. The genomic services system of claim 1 , wherein:
the predetermined threshold probability is 1%;
the error probabilities of the high-quality subset of genotypes do not exceed 1%; and
prior to modification by the first imputation, the error probabilities of the low-quality subset of genotypes exceed 1%.
6. The genomic services system of claim 1 , wherein the operations further comprise:
between the generating of the pre-imputation file and the first imputation, phasing the high-quality subset of genotypes without phasing the low-quality subset of genotypes; and
between the first imputation and the second imputation, phasing both the high-quality subset of genotypes and the low-quality subset of genotypes.
7. A method comprising:
receiving, by a network interface, genomic sequence reads of a user, the genomic sequence reads of the user having been generated by genomic sequencing equipment based on a biological sample provided by the user;
aligning, by one or more processors of a machine, the received genomic sequence reads of the user to a reference human genome;
identifying, by one or more processors of the machine, genomic variants of the user based on a comparison of the aligned sequence reads of the user to the reference human genome;
generating, by one or more processors of the machine, a pre-imputation file that corresponds to the user and specifies genotypes of the user, the pre-imputation file specifying the genotypes of the user by specifying the identified genomic variants of the user, the pre-imputation file indicating a high-quality subset of genotypes whose error probabilities do not transgress a predetermined threshold probability, the pre-imputation file indicating a low-quality subset of genotypes whose error probabilities transgress the predetermined threshold probability;
performing, by one or more processors of the machine, a first imputation that adds first additional genotypes to the genotypes specified in the pre-imputation file based on the high-quality subset of genotypes and without influence from the low-quality subset of genotypes, the first imputation modifying at least some of the error probabilities of the low-quality subset of genotypes;
performing, by one or more processors of the machine, a second imputation that adds second additional genotypes to the genotypes specified in the pre-imputation file based on both the high-quality subset of genotypes whose error probabilities do not transgress the predetermined threshold probability and the low-quality subset of genotypes at least some of whose error probabilities were modified by the performed first imputation; and
generating, by one or more processors of the machine, a post-imputation file that corresponds to the user, specifies the genotypes specified in the pre-imputation file, and further specifies the first and second additional genotypes of the user, based on the performed first and second imputations.
8. The method of claim 7 , further comprising:
determining error probabilities of the genotypes specified by the pre-imputation file, the determined error probabilities including the error probabilities of the high-quality subset of genotypes and the error probabilities of the low-quality subset of genotypes.
9. The method of claim 8 , wherein:
variations in the assessed error probabilities of the genotypes specified by the pre-imputation file indicate variations in depth of coverage among the genomic sequence reads generated by the genomic sequencing equipment.
10. The method of claim 7 , wherein:
the first imputation is included in a first phase of a dual-phase imputation in which the first phase provides results to a second phase of the dual-phase imputation; and
the second imputation is included in the second phase of the dual-phase imputation and performed based on the results provided by the first phase of the dual-phase imputation.
11. The method of claim 7 , wherein:
the predetermined threshold probability is 1%;
the error probabilities of the high-quality subset of genotypes do not exceed 1%; and
prior to modification by the first imputation, the error probabilities of the low-quality subset of genotypes exceed 1%.
12. The method of claim 7 , further comprising:
between the generating of the pre-imputation file and the first imputation, phasing the high-quality subset of genotypes without phasing the low-quality subset of genotypes; and
between the first imputation and the second imputation, phasing both the high-quality subset of genotypes and the low-quality subset of genotypes.
13. A non-transitory machine-readable medium comprising instructions that, when executed by one or more processors of a machine, cause the machine to perform operations comprising:
receiving genomic sequence reads of a user, the genomic sequence reads of the user having been generated by genomic sequencing equipment based on a biological sample provided by the user;
aligning the received genomic sequence reads of the user to a reference human genome;
identifying genomic variants of the user based on a comparison of the aligned sequence reads of the user to the reference human genome;
generating a pre-imputation file that corresponds to the user and specifies genotypes of the user, the pre-imputation file specifying the genotypes of the user by specifying the identified genomic variants of the user, the pre-imputation file indicating a high-quality subset of genotypes whose error probabilities do not transgress a predetermined threshold probability, the pre-imputation file indicating a low-quality subset of genotypes whose error probabilities transgress the predetermined threshold probability;
performing a first imputation that adds first additional genotypes to the genotypes specified in the pre-imputation file based on the high-quality subset of genotypes and without influence from the low-quality subset of genotypes, the first imputation modifying at least some of the error probabilities of the low-quality subset of genotypes;
performing a second imputation that adds second additional genotypes to the genotypes specified in the pre-imputation file based on both the high-quality subset of genotypes whose error probabilities do not transgress the predetermined threshold probability and the low-quality subset of genotypes at least some of whose error probabilities were modified by the performed first imputation; and
generating a post-imputation file that corresponds to the user, specifies the genotypes specified in the pre-imputation file, and further specifies the first and second additional genotypes of the user, based on the performed first and second imputations.
14. The medium of claim 13 , wherein the operations further comprise:
determining error probabilities of the genotypes specified by the pre-imputation file, the determined error probabilities including the error probabilities of the high-quality subset of genotypes and the error probabilities of the low-quality subset of genotypes.
15. The medium of claim 14 , wherein:
variations in the assessed error probabilities of the genotypes specified by the pre-imputation file indicate variations in depth of coverage among the genomic sequence reads generated by the genomic sequencing equipment.
16. The medium of claim 15 , wherein:
the first imputation is included in a first phase of a dual-phase imputation in which the first phase provides results to a second phase of the dual-phase imputation; and
the second imputation is included in the second phase of the dual-phase imputation and performed based on the results provided by the first phase of the dual-phase imputation.
17. The medium of claim 16 , wherein:
the predetermined threshold probability is 1%;
the error probabilities of the high-quality subset of genotypes do not exceed 1%; and
prior to modification by the first imputation, the error probabilities of the low-quality subset of genotypes exceed 1%.
18. The medium of claim 17 , wherein the operations further comprise:
between the generating of the pre-imputation file and the first imputation, phasing the high-quality subset of genotypes without phasing the low-quality subset of genotypes; and
between the first imputation and the second imputation, phasing both the high-quality subset of genotypes and the low-quality subset of genotypes.