haplotypes in the cell panel (
Figure 6
and
Table S3
). Consis-
tent with our results are KIR genotyping data obtained pre-
viously from the IHWG cells, which achieved a limited
discrimination of alleles.
6,59
These analyses demonstrate
that PING accurately processes high-throughput sequence
data to give accurate high-resolution KIR genotypes. The
high-resolution KIR genotypes of the IHWG cell panel
have been compiled (
Table S3
), providing a resource for
future investigation.
Validation of the High-Resolution HLA Class I Capture
Method
Because KIR and HLA class I glycoproteins are function-
ally interacting receptors and ligands, the capture/NGS
method was designed to capture and analyze the two
gene families in the same reaction (
Material and Methods
).
To validate the HLA class I sequences obtained by this
method, we first analyzed the panel of 97 IHWG cell lines.
In previous studies,
38–40
including high-density whole-
genome SNP analysis,
43
90 of the IHGW cells were judged
to be homozygous for HLA-A, HLA-B, and HLA-C, and five
of the other seven lines were found to be homozygous
for two of the three HLA class I genes. Our high-
throughput sequencing results are completely concordant
with the genotypes previously determined by conven-
tional methods
15
(
Table S4
A). In the course of validation,
we defined two novel HLA-C alleles that encode distinctive
proteins (
Table S4
B). Because our method gives full-length
genomic sequences, including introns and flanking re-
gions, all HLA-A, HLA-B, and HLA-C alleles of the IHWG
cells are now defined at a much higher resolution (‘‘four
field’’; see
Appendix A
) than previously achieved.
The IHWG cells represent an unusual, and highly
selected, sampling of the human population because
they are HLA homozygous, and many of them derive
from consanguineous individuals. To extend the study to
heterozygous individuals, we applied the capture/NGS
method to 30 West African family trios from Mali (sample
set 2,
Material and Methods
) and 188 individuals selected
at random from a panel of healthy Europeans (sample
set 3). The capture/NGS method achieved full coverage of
the HLA class I genes, and an example of the result ob-
tained from one individual is shown in
Figure 7
A. In the
Africans, 22 HLA-A, 30 HLA-B, and 15 HLA-C alleles were
identified. These allele sequences agreed completely with
HLA class I genotypes we determined by standard probe-
based and Sanger sequencing methods. They were also
consistent with the observed segregation of HLA-A,
HLA-B, and HLA-C alleles within the family trios (
Tables
S4C–S4
D). In total, 170 distinct HLA class I alleles were
identified in the three validation sets (IHWG cell lines, Af-
rican trios, and Europeans). These 170 alleles include 62 of
the 69 fundamental allele types defined by the first two
digits of the HLA nomenclature (
Figure 7
B) and thus cover
most, if not all, of the breadth of HLA class I allelic diver-
sity.
21
Thus, it is likely that all HLA-A, HLA-B, and HLA-C
alleles can be captured and sequenced by our method. In
support of this thesis, the capture/NGS method robustly
detects and sequences B*73:01 (E.A., unpublished data), a
unique, archaic allele that has by far the most divergent
KIR
1. SP0010
2. CB6B
3. E481324
4. LZL
3DL3
*00206
*00206
*01403
*00301
*00206
*01501
*00802
*00402
2DS2
*001
*001
*001
2DL2/3
L3*001
L3*001
L2*001
L2*003
L3*001
L3*001
L3*001
L2*003
2DL5
B*002
2DS3/5
3*00103
2DP1
*00203
*00203
*012
*00203
*00204
*002
2DL1
*00302
*00302
*00401
*00302
*00302
*00302
3DP1
*00302
*00302
*00301
*009
*00302
*00302
*00302
3DP1b
*001
2DL4
*011
*011
*00501
*00501
*00103
*00802
*00102
2DL4b
*00501
3DL1/S1
L1*00501 L1*00501
S1*013
S1*013
L1*002
L1*00401
L1*01502
3DL1/S1b
3DS1*013
2DL5
A*001
A*005
B*010
2DS3/5
5*002
3*002
3*002
2DS1
*002
*002
*002
2DS4
*010
*010
*00101
*006
*00101
3DL2
*00103
*00103
*00701
*00701
*00103
*00501
*00201
*00701
Figure 6.
High-Resolution KIR Allele and Copy-Number Genotypes of 97 IHWG Cells
Four examples of high-resolution allele and copy-number genotypes of KIR. Individual 1 (SP0010) is homozygous for the KIR A haplo-
type. Individual 2 (CB6B) has two different B haplotypes. Individual 3 (E481324) has a duplication of three loci (in blue shading: denoted
as 3DP1b, 2DL4b, and 3DL1/S1b). Individual 4 (LZL) has a deletion of the central segment of the KIR haplotype (red). Yellow shading
denotes alleles that were identified in the current study. The full genotypes for each IHWG cell are given in
Table S3
.
384
The American Journal of Human Genetics 99, 375–391, August 4, 2016