Bioinformatics-2011-Cibulskis-bioinformatics-btr446(1).pdf

(324 KB) Pobierz
Bioinformatics Advance Access published July 29, 2011
ContEst: Estimating cross-contamination of human samples in
next generation sequencing data
Kristian Cibulskis
1,*,+
, Aaron McKenna
1,+
, Tim Fennell
1
, Eric Banks
1
, Mark DePristo
1
and Gad
Getz
1,*
1
Broad Institute, 7 Cambridge Center, Cambridge MA 02142
Associate Editor: Prof. Martin Bishop
Downloaded from
http://bioinformatics.oxfordjournals.org/
at Politechnika Poznanska on November 22, 2015
ABSTRACT
Summary:
Here, we present
ContEst,
a tool for estimating the level
of cross-individual contamination in next generation sequencing
data. We demonstrate the accuracy of ContEst across a range of
contamination levels, sources, and read depths using sequencing
data mixed in-silico at known concentrations. We applied our tool to
published cancer sequencing data sets and report their estimated
contamination levels.
Availability and Implementation:
ContEst is a GATK (McKenna, et
al., 2010) module, and distributed under a BSD style license at
http://www.broadinstitute.org/cancer/cga/contest
Contact:
kcibul@broadinstitute.org , gadgetz@broadinstitute.org
Supplementary information:
Supplementary data is available at
Bioinformatics online
ity. Cross-species contamination is easily detected by aligning to
unique regions of potentially contaminating species. In order to
address the most critical need, we developed ContEst to accurately
estimate the cross-individual contamination level in next genera-
tion sequencing data.
2
METHODS
Given genotype information about the sequenced sample from a genotyp-
ing array in VCF format (http://www.1000genomes.org), general popula-
tion frequency information (provided with ContEst), and the sequencing
data in BAM format (Li, et al., 2009), we use a Bayesian approach to calcu-
late the posterior probability of the contamination level and determine the
maximum a posteriori probability (MAP) estimate of the contamination
level.
The method first identifies the homozygous SNP sites based on the array
data,
S≡{s
i
},
i=1,…,N,
and the alleles at these sites,
A≡{A
i
}. For each site,
s
i
, we denote the probability in the contaminating population to observe
A
i
at that site by
f
i
, and therefore the probability to see the other allele is 1-
f
i
.
In addition, we denote by
b
ij
and
e
ij
the called base of the
j-th
read that
covers
s
i
and its quality (represented by its probability of being incorrect),
respectively. The number of reads that cover
s
i
, i.e. the depth at that site, is
denoted by
d
i
. For a contamination fraction c, we can now calculate the
posterior probability using the Bayes rule:
P(c
|
B, E, A, F)
=
P(B
|
c, E, A, F)P(c)
P(B)
1
INTRODUCTION
Next generation sequencing methods are generating vast amounts
of short sequence reads for the purpose of studying DNA sequence
variations, and identifying those that affect human disease. Many
novel methods allow for the interrogation of the structure of the
genome with unprecedented sensitivity due to the digital nature of
the data (Trapnell and Salzberg, 2009). Rare events present in only
a fraction of the sequenced material, as is the case in somatic muta-
tion discovery in cancer genome studies (Chapman, et al., 2011)
(Berger, et al., 2011), can be accurately detected by sequencing to
greater read depth. Moreover, genome partitioning techniques
(Gnirke, et al., 2009) allow for even greater sensitivity at a lower
cost by targeting only regions of interest.
However, these methods can be heavily compromised by con-
tamination. Three major classes of DNA contamination exist:
cross-individual, within-individual, and cross-species. Cross-
individual is the most critical to control, as even small levels of
contamination can cause many false positives, particularly in con-
trastive tumor vs. normal cancer studies (Fig. 1a). Within-
individual contamination, such as normal DNA contamination of
tumor DNA in cancer studies, typically leads to decreased sensitiv-
*
To
+
Using a uniform prior on
c, i.e.
P(c)=1, and assuming that the reads (and
noise) are independent and equivalent for all 3 types of substitutions and
discarding sites suspected to be genotyping array data errors (see Supple-
mental Methods), we obtain:
P(c
|
B, E, A, F)
∝
P(B
|
c,E, A, F)
=
∏ ∏
P(b
ij
|
e
ij,
A
i
,
f
i
)
i=1 j
=1
N
d
i
Where
(1
−
c)(1
−
e
)
+c 
f
(1
−
e
)
+
(1
−
f
)(e / 3)
if b = A
ij
ij
i
ij
ij
i
i
P(b
ij
|
e
ij
,
A
i
,
f
i
)
=
(1
−
c)(e
ij
/ 3)
+c 
f
i
(e
ij
/ 3)
+
(1
−
f
i
)(1
−
e
ij
)
if b
ij
= A
i
e
ij
/ 3
otherwise
whom correspondence should be addressed.
The authors wish it to be known that, in their opinion, the first two authors
should be regarded as joint First authors
The qualities of bases are typically represented using a Phred-like Q-
scores, i.e. e=10
–q/10
. Finally, we evaluate the above equation for
c
∈
[0, 1]
and normalize to 1 in order to get the posterior probability. The MAP esti-
mate of
c
is the mode of this distribution, and a confidence interval can be
© The Author (2011). Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oup.com
1
Downloaded from
http://bioinformatics.oxfordjournals.org/
at Politechnika Poznanska on November 22, 2015
Fig. 1 (A) False-positive somatic mutations detected per megabase on in-silico contaminated data; most cancers have ~1 true event per megabase (B) accuracy with single con-
taminating sample (C) accuracy with multiple contaminating samples (D) accuracy with respect to read depth; shaded areas indicate 95% confidence interval (E) contamination
estimates of TCGA Ovarian dataset
calculated using the minimal interval containing 95% of the posterior prob-
ability. Note that reads that do not support a known allele at
S
contribute a
factor that is independent of
c,
hence we can ignore them in the calculation.
For tumor samples, we recommend using the genotypes of the patient-
matched normal when available instead of the tumor, since homozygous
SNPs in regions of loss-of-heterozygosity in the tumor will interpret con-
tamination with normal cells from the same patient as foreign DNA since
they have different genotypes.
is a significant fraction of the typical somatic mutation rate of
1/mb per sample.
In addition, ContEst has proven to be essential in lab quality
control to identify and monitor sources of contamination, which
has helped decrease contamination at the Broad Institute.
ACKNOWLEDGEMENTS
We would like to acknowledge our colleagues from the Broad
Sequencing Platform, Genetics Analysis Platform, and The Cancer
Genome Atlas Project who supported the development of ContEst,
as well as Rameen Beroukhim for valuable discussions.
Funding:
This work was supported by the National Human Ge-
nome Research Institute [grant number U24 CA126546].
3
RESULTS
Using next generation sequencing data from the TCGA Ovarian
publication (TCGA Research Network, 2011), we identified 12
exome-capture BAMs with low contamination, having very few
reads that do not match the homozygous calls from their genotyp-
ing arrays (Supp. Table 1). Next, we created in-silico data sets by
mixing a primary sample with one or more contaminants at spe-
cific contamination levels (See Supp. Material). Reassuringly, the
estimate of the contamination level of the primary sample alone
was 0.08%. ContEst was able to accurately predict the level of
contamination across a wide range of conditions including more
than a single contaminating sample. (Fig 1b,c)
In order to assess the accuracy as a function of sequencing depth
we down-sampled the depth of the sequencing data (Fig 1d), and
demonstrated that ContEst produces accurate estimates even with
average coverage < 5x.
Applying the method to data obtained from the TCGA Ovarian
publication (Supp. Table 2) indicates that low levels of physical
contamination are common (Fig 1e). Independent validation of all
somatic events likely ensured that this contamination did not cause
false positives in the publication. However, given a distribution of
contamination as seen in TCGA (Fig 1e), and an estimated error
rate at non-dbSNP sites from contamination as shown in Figure 1a,
a typical cancer project might expect > 10% of the samples to have
> 1.5% contamination, causing ~0.2 errors/mb per sample, which
REFERENCES
Berger, M.F., et al. (2011) The genomic complexity of primary
human prostate cancer, Nature, 470, 214-220.
Chapman, M.A., et al. (2011) Initial genome sequencing and
analysis of multiple myeloma, Nature, 471, 467-472.
Gnirke, A., et al. (2009) Solution hybrid selection with ultra-long
oligonucleotides for massively parallel targeted sequencing, Nat
Biotechnol, 27, 182-189.
Li, H., et al. (2009) The Sequence Alignment/Map format and
SAMtools, Bioinformatics, 25, 2078-2079.
McKenna, A., et al. (2010) The Genome Analysis Toolkit: a
MapReduce framework for analyzing next-generation DNA
sequencing data, Genome Res, 20, 1297-1303.
TCGA Research Network (2011) Integrated Genomic Analyses of
Ovarian Carcinoma Nature, 474, 609-615.
Trapnell, C. and Salzberg, S.L. (2009) How to map billions of
short reads onto genomes, Nat Biotechnol, 27, 455-457.
2

Plik z chomika:

Inne pliki z tego folderu:

Inne foldery tego chomika:

Zgłoś jeśli naruszono regulamin