Systems and methods for tumor fraction estimation from small variants
Systems and methods for cancer subject tumor fraction estimation comprise obtaining a first plurality of nucleic acid fragment sequences from the subject's liquid biological sample. The first plurality of sequences represent cell-free nucleic acids in the liquid sample. A second plurality of nucleic acid fragment sequences is obtained from the subject's tumor sample. The second plurality of sequences represent nucleic acid molecules in the tumor. Smoothed noise rates, each determined using nucleic acid fragment sequences from non-cancer samples mapping to a corresponding allele position in a plurality of allele positions, are obtained. Variant allele counts and coverages are determined for the allele positions using the first plurality of sequences. Solid variant allele fractions are determined for the plurality of allele positions using the second plurality of sequences. The subject tumor fraction is calculated using the smoothed noise rates, variant allele counts, coverages, and solid variant allele fractions.
1 . A method of estimating a tumor fraction f of a subject, the method comprising:
(A) sequencing a first plurality of nucleic acid fragment sequences from a liquid biological sample obtained from the subject, wherein the liquid biological sample comprises blood, whole blood, plasma, serum, urine, fecal, saliva, sweat, or tears obtained from the subject;
(B) sequencing a second plurality of nucleic acid fragment sequences from a solid tumor sample obtained from the subject, wherein an average coverage rate for at least one of the first plurality of nucleic acid fragment sequences and the second plurality of nucleic acid fragment sequences is less than 5000×;
(C) determining a plurality of estimated smoothed noise rates corresponding to a plurality of allele positions, wherein the plurality of allele positions comprise a range between one million to ten million allele positions, each respective estimated smoothed noise rate ev i in the plurality of estimated smoothed noise rates is determined using nucleic acid fragment sequences, in electronic form, obtained from each non-cancer sample in a cohort of non-cancer samples mapping to a corresponding allele position v i in the plurality of allele positions, and each respective estimated smoothed noise rate ev i for the corresponding allele position specifies how often a tumor allele appears at the corresponding allele position in non-cancer samples;
(D) determining a variant allele count s vi,t for each respective allele position v; in the plurality of allele positions using the first plurality of nucleic acid fragment sequences, thereby obtaining a plurality of variant allele counts for the subject;
(E) determining a coverage s vi for each respective allele position v i in the plurality of allele positions using the first plurality of nucleic acid fragment sequences, thereby obtaining a plurality of coverages for the subject;
(F) determining a solid variant allele fraction a; for each respective allele position v i in the plurality of allele positions using the second plurality of nucleic acid fragment sequences, thereby obtaining a plurality of solid variant allele fractions;
(G) determining, using a computer system comprising a processor coupled to non-transitory memory, the tumor fraction f of the subject based on D s comprising (i) the plurality of estimated smoothed noise rates, (ii) the plurality of variant allele counts, (iii) the plurality of coverages, and (iv) the plurality of solid variant allele fractions, wherein (ii) and (iii) are determined based on the first plurality of nucleic acid fragment sequences from the liquid biological sample, and (iv) is determined based on the second plurality of nucleic acid fragment sequences from the solid tumor sample; and
(H) modifying a treatment regimen for the subject based, at least in part, on a value of the tumor fraction f of the subject, wherein the treatment regimen comprises applying an agent for cancer to the subject, wherein the agent for cancer is a hormone, an immune therapy, radiography, or a cancer drug, and wherein the modification to the treatment regimen comprises intensifying or discontinuing the agent.
2 . The method of claim 1 , wherein the respective estimated smoothed noise rate ev i for a respective allele position v i is calculated as:
e
v
i
=
(
n
v
i
,
t
+
b
)
(
n
v
i
+
2
b
)
wherein,
n v i ,t is a variant allele count at position v i over the cohort of non-cancer samples,
n v i is coverage of nucleic acid fragment sequences at v i summed over the cohort of non-cancer samples, and
b is a positive value representing a pseudocount.
3 . The method of claim 1 , wherein the determining (G) calculates a likelihood that a candidate tumor fraction f c is a true tumor fraction of the subject in the form of a posterior distribution of the candidate tumor fraction f c given D s by evaluating:
P ( D s )∝ P ( f c ) P ( f c ),
wherein,
P(f c ) is a prior, wherein the prior is a non-informative prior,
P(D s ) is the posterior distribution of the candidate tumor fraction f c given D s , and
P(f c ) is a probability of observing D s given the tumor fraction f c .
4 . The method of claim 3 , wherein the determining (G) comprises computing the tumor fraction f of the subject by computing D s as part of a grid search, a Markov chain Monte Carlo sampling, or an Expectation-Maximization algorithm.
5 . The method of claim 3 , wherein the tumor fraction of the subject is deemed to be a median value of P(D s ) across a range of calculated tumor fractions.
6 . The method of claim 3 , wherein the posterior distribution assumes a set of information comprising a plurality of solid variant allele fractions, wherein the posterior distribution is calculated based on a conditional probability of the set of information being satisfied, assuming that the candidate tumor fraction satisfies f c .
7 . The method of claim 3 , wherein the determining the tumor fraction f in (G) comprises calculating:
wherein,
s v i ,t is a variant allele count at allele position v i ,
s v i is a coverage for s v i ,t at allele position v i , and
a i is a solid variant allele fraction at allele position v i .
8 . The method of claim 3 , wherein the determining the tumor fraction f in (G) comprises calculating:
wherein,
is normalized,
x v i =s v i ,t ,
n v i =s v i ,
P v i =f c a i +(1−f c )ev i , and
a i =the solid variant allele fraction at allele position v i .
9 . The method of claim 3 , wherein the calculation of the tumor fraction f in (G) comprises a factor corresponding to a germline contamination:
wherein,
s v i ,t is a variant allele count at allele position v i ,
s v i is a coverage for s v i ,t at allele position v i ,
a i is a solid variant allele fraction at allele position v i ,
Q and P are each positive and sum to one, and
T is between 0.45 and 0.55.
10 . The method of claim 1 , the method further comprising calculating a threshold credible interval for the tumor fraction f of the subject.
11 . The method of claim 1 , the method further comprising using at least the tumor fraction f of the subject to determine a cancer condition of the subject.
12 . The method of claim 1 , wherein the determining (D) comprises determining a number of nucleic acid fragment sequences in the first plurality of nucleic acid fragment sequences having a respective variant at each respective allele position in the plurality of allele positions.
13 . The method of claim 12 , wherein the determining (D) comprises aligning a nucleic acid fragment sequence in the first plurality of nucleic acid fragment sequences to a region in a reference genome in order to determine whether the nucleic acid fragment sequence includes the respective variant.
14 . The method of claim 1 , wherein the determining the solid variant allele fraction a i for each respective allele position v i in the plurality of allele positions using the second plurality of nucleic acid fragment sequences (F) comprises comparing a number of nucleic acid fragment sequences in the second plurality of nucleic acid fragment sequences having a respective variant at each respective allele position in the plurality of allele positions to a total number of nucleic acid fragment sequences in the second plurality of nucleic acid fragment sequences that map to the respective allele position.
15 . The method of claim 14 , wherein the determining (F) comprises aligning a nucleic acid fragment sequence in the second plurality of nucleic acid fragment sequences to a region in a reference genome in order to determine whether the nucleic acid fragment sequence includes the respective variant.
16 . The method of claim 1 , wherein the plurality of allele positions are selected for a particular type of cancer, or
the plurality of allele positions are selected for a plurality of cancers, or
the plurality of allele positions are selected independent of a type of cancer.
17 . The method of claim 1 , wherein the plurality of allele positions comprises a methylation pattern allele position.
18 . The method of claim 1 , wherein at least one of the first plurality of nucleic acid fragment sequences and the second plurality of nucleic acid fragment sequences is obtained using whole-genome bisulfite sequencing.
19 . The method of claim 1 , further comprising evaluating a stage of the cancer in the subject based on the calculated tumor fraction f of the subject.
20 . The method of claim 1 , further comprising evaluating a level of aggressiveness of the cancer in the subject based on the calculated tumor fraction f of the subject.
21 . The method of claim 1 , wherein the agent for cancer is Lenalidomide, Pembrolizumab, Trastuzumab, Bevacizumab, Rituximab, Ibrutinib, Human Papillomavirus Quadrivalent (Types 6, 11, 16, and 18) Vaccine, Pertuzumab, Pemetrexed, Nilotinib, Denosumab, Abiraterone acetate, Promacta, Imatinib, Everolimus, Palbociclib, Erlotinib, or Bortezomib.
22 . The method of claim 1 , wherein the subject has been treated with the agent for cancer and the method further comprises:
(I) using the tumor fraction f of the subject to evaluate a response of the subject to the agent for cancer.