Bambam: parallel comparative analysis of high-throughput sequencing data
The present invention relates to methods for evaluating and/or predicting the outcome of a clinical condition, such as cancer, metastasis, AIDS, autism, Alzheimer's, and/or Parkinson's disorder. The methods can also be used to monitor and track changes in a patient's DNA and/or RNA during and following a clinical treatment regime. The methods may also be used to evaluate protein and/or metabolite levels that correlate with such clinical conditions. The methods are also of use to ascertain the probability outcome for a patient's particular prognosis.
1. A parallel genomic comparative analysis system comprising:
a memory; and
an sequence analysis engine coupled with the memory and configured to:
identify a genomic position within a reference genome;
access a first file storing tumor sequence data including short reads associated with a tumor tissue;
access a second file storing match normal sequence data short reads associated with a matched normal tissue;
store in the memory a tumor dataset having tumor short read sequences from the first file where the tumor short read sequences overlap the genomic position;
store in the memory a matched normal dataset having matched normal short read sequences from the second file and that overlap the genomic position;
select a tumor genotype and a matched normal genotype that maximize a joint probability as a function of the tumor short read sequences and the match normal short read sequences at the genomic position, wherein the joint probability depends on one of a probability calculated as a multinomial operating as a function of the matched normal genotype or as a probability calculated as a multinomial operating as a function of the tumor genotype; and
store a difference between the tumor genotype and the matched normal genotype in a device memory.
2. The system of claim 1 , wherein the memory is configured to store the tumor dataset and the matched normal dataset simultaneously.
3. The system of claim 1 , wherein the sequence analysis engine is further configured to synchronize the first and the second file.
4. The system of claim 1 , wherein the sequence analysis engine is further configured to read the first and the second files at the same time to access the files.
5. The system of claim 1 , wherein the genomic position comprises a common genomic location between the first and the second file relative to the reference genome.
6. The system of claim 1 , wherein the difference is selected from the group consisting of:
a somatic variant, a germline variant, a single nucleotide polymorphism, an allele-specific copy number, a loss of heterozygosity, a structural rearrangement, a chromosomal fusion, and a breakpoint.
7. The system of claim 1 , wherein the analysis engine is further configured to calculate a confidence score of the tumor genotype and the match normal genotype pair.
8. The system of claim 7 , wherein the analysis engine is configured to calculate the confidence scores as a posterior probability.
9. The system of claim 7 , wherein the analysis engine is further configured to store the confidence score in the device memory with the difference.
10. The system of claim 1 , wherein at least one of the first file and the second file comprises at least one of a BAM file and a SAM file.
11. The system of claim 1 , wherein the tumor dataset comprises all tumor short read sequences in the first file that overlap the genomic position.
12. The system of claim 1 , wherein the matched normal dataset comprises all matched normal short read sequences in the second file that overlap the genomic position.
13. The system of claim 1 , wherein the tumor sequence data of the tumor tissue and the matched normal sequence data of the matched normal tissue originate from the same person.
14. A parallel genomic comparative analysis
system comprising: a memory; and
a sequence analysis engine coupled with the memory and configured to:
access a first file storing tumor sequence data including short reads associated with a tumor tissue;
access a second file storing matched normal sequence data short reads associated with a matched normal tissue;
align, relative to a first genomic position within a reference genome, the short reads associated with the tumor tissue with the short reads associated with the matched normal tissue;
process at the same time all aligned short reads to determine a difference between the tumor sequence data and the matched normal sequence data;
store a difference between the tumor sequence data and the matched normal sequence data in a device memory;
align, relative to a second genomic position within the reference genome, the short reads associated with the tumor tissue with the short reads associated with the matched normal tissue;
process at the same time all aligned short reads to determine a second difference between the tumor sequence data and the matched normal sequence data; and
store the second difference between the tumor sequence data and the matched normal sequence data in the device memory.
15. The system of claim 14 wherein the sequence analysis engine is further configured to select a tumor genotype and a matched normal genotype that maximize a joint probability as a function of the tumor short read sequences and the match normal short read sequences at the genomic position.
16. The system of claim 15 wherein maximizing the joint probability is defined by
P ( D r ,Ac,r )−− P ( IVG ,) P ( G,Ir′GA .;), P (1)
D 4 , , , G; LIF G 7 Ar )*( DG 4 ,G 4 , , , LO (9 10 . . . − i 6; 48 ) (2)
where r is an observed reference allele, a is a fraction of normal contamination, where tumor and germline genotypes are defined by Gt=(ti, t 2 ) and Gg=(gi,g2), where t 1 , t2, gi, g2 c {A, T, C, G}, and where tumor and germline sequence data are defined as a set of reads DM d i 1 , d i 2 . . . , d i n } and D g ={d g i , d g 2 . . . , d g n }, respectively, with the observed bases d i ′, d g ′E {A, T, C, G}.
17. The system of claim 15 wherein a probability of germline alleles given a germline genotype is modeled as a multinomial over four nucleotides:
where n is the total number of germline reads at the genomic position and n A , n G , nc, n t are reads supporting each observed allele, and a probability of tumor alleles given a tumor genotype is modeled as a multinomial over four nucleotides:
P
(
D
i
❘
D
t
,
G
g
,
α
)
=
n
!
n
A
!
n
T
!
n
G
!
n
C
!
∏
i
n
P
(
d
g
i
❘
G
t
,
G
g
,
α
)
,
where n is the total number of germline reads at the genomic position and n A , n G , n c , n t are reads supporting each observed allele.
18. A parallel genomic comparative analysis system comprising:
a memory; and
an sequence analysis engine coupled with the memory and configured to:
identify a genomic position within a reference genome;
access a first file storing tumor sequence data including short reads associated with a tumor tissue;
access a second file storing match normal sequence data short reads associated with a matched normal tissue;
store in the memory a tumor dataset having tumor short read sequences from the first file where the tumor short read sequences overlap the genomic position, wherein the tumor dataset comprises all tumor short read sequences in the first file that overlap the genomic position;
store in the memory a matched normal dataset having matched normal short read sequences from the second file and that overlap the genomic position;
select a tumor genotype and a matched normal genotype that maximize a joint probability as a function of the tumor short read sequences and the match normal short read sequences at the genomic position; and
store a difference between the tumor genotype and the matched normal genotype in a device memory.
19. The system of claim 18 , wherein the matched normal dataset comprises all matched normal short read sequences in the second file that overlap the genomic position.