NMETH-BC07899D
Supplementary Figure 1. The principle of 7DDNA haplotyping. A diploid cell is shown
with 5 chromosome pairs. By computer-directed laser microdissection at the dotted line, 5
chromosomes are collected. In this harvest, chromosomes 2, 3 and 5 are monosomic,
chromosome 1 is disomic, and chromosome 4 is null. Conventional genotyping using this
harvest will directly reveal haplotypes of chromosomes 2, 3, and 5.
NMETH-BC07899D
Supplementary Figure 2. Monitoring the copy
number of chromosome-1 in a microdissection
harvest. The copy number of each sister
chromosome pair in the harvest can be readily
monitored by the genotypes of the 7DDNA sample. If
a 7DDNA sample shows “homozygous” genotypes at
all loci known to be heterozygous (known from
genomic DNA genotyping results) along a
chromosome in a person, this microdissection
harvest should contain one single copy of this
chromosome. This figure shows observed
heterozygous SNPs along the chromosome-1 of our
microdissection harvest-2 in our genotyping result
using the Illumina CNV370 platform. Each red bar
represents a known heterozygous SNP that remained
to be heterozygous in the genotype call with WGA4
amplified microdissection harvest. Those SNPs
located on repeats were excluded. These observed
heterozygous SNP loci enabled us to distinguish the
haploid portion (containing only one of two sister
chromosome copies) and the diploid portion
(containing both sister chromosome copies) in this
microdissected sample. The boundary between the
haploid portion and the diploid portion is indicated by
a horizontal green line in this figure. One of the
homologous copies of chromosome-1 was broken in
the microdissection procedure, the allele calls of
chromsome-1 segment below this line constituted the
haplotype output.
NMETH-BC07899D
Supplementary Figure 3. Monitoring the copy number of chromosomes in three
microdissection harvests. The copy number of each sister chromosome pair in a
microdissection harvest can be monitored through the remaining heterozygous loci in a
sample. If a 7DDNA sample shows “homozygous” genotypes at all loci known to be
heterozygous (known from genomic DNA genotyping results) along a chromosome in a
person, this microdissection harvest should contain one single copy of this chromosome.
This figure shows two typical windows (one for haploid chromosomes and one for diploid
chromosomes) for each of our three microdissection harvests. Each red bar represents a
known heterozygous SNP that remained to be heterozygous in the genotype call with
WGA4 amplified microdissection harvest. Those SNPs located on repeats were excluded.
NMETH-BC07899D
Supplementary Figure 4. The relationship between chromosome coverage and the
cumulative number of microdissection in the 7DDNA haplotyping. For example, 6
cumulative cut will ensure ~72.9% of the total individuals receive a whole-genome
haplotype report , in which 20.1% of individuals may have completed their allchromosome haplotyping when it reaches 4 cuts, 26.2% of individuals may have
completed in the 5th round, and 20.7% of individuals at the 6th round. The probability data
is provided in Supplementary Table 6. The detailed calculation procedure is provided in
Supplementary Equation-1.
NMETH-BC07899D
Supplementary Table 1. A zoom-in window of the haplotype report on chromosome 5.
SNP
Name
rs10076494
rs1108867
rs28416084
rs6869925
rs10055993
rs4957023
rs6868535
rs4541696
rs4045344
rs9313132
rs7702501
rs920980
rs4957112
rs4594899
rs4975605
rs2736100
rs402710
rs31489
rs6347
rs27048
Chr
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
Pos
(bp)
91139
222902
270640
317036
317296
389952
548377
634053
648870
651237
918411
949983
1E+06
1E+06
1E+06
1E+06
1E+06
1E+06
1E+06
1E+06
SNP
[A/G]
[A/G]
[A/G]
[A/G]
[A/C]
[T/C]
[T/C]
[A/G]
[A/C]
[A/G]
[A/C]
[T/C]
[T/C]
[A/G]
[T/G]
[T/G]
[T/C]
[A/C]
[A/G]
[T/C]
Conventional
genotyping output
gDNA
7DDNA
TC
AG
AG
AG
AC
TC
TC
TC
AC
AG
AC
TC
TC
AG
AC
TG
TC
AC
AG
TC
Haplotype
output
Observed Derived
TT
AA
AA
GG
CC
TT
CC
TT
AA
AA
AA
CC
CC
GG
CC
TT
TT
AA
GG
CC
T
A
A
G
C
T
C
T
A
A
A
C
C
G
C
T
T
A
G
C
C
G
G
A
A
C
T
C
C
G
C
T
T
A
A
G
C
C
A
T
One chromosomal haplotype (observed) was directly readout from the Illumina output, the
other chromosomal haplotype (derived) is composed of opposite alleles that were not
observed in the genotyping report. gDNA, genomic DNA. Conventional genotyping, we used
a high-throughput Illumina genotyping platform in this experiment.
NMETH-BC07899D
Supplementary Table 3. Fidelity of whole genome amplification in the 7DDNA haplotyping.
Error rate (%)
Samples
haploid chromosomes
diploid chromosomes
all chromosomes
Sample-1
1.66
1.60
1.63
Sample-2
0.63
0.50
0.54
Sample-3
0.72
0.46
0.50
All samples
0.98
0.64
0.73
Through observation on the genotype calls of known homozygous SNPs after whole genome
amplification of our microdissection harvests, we were able to access the fidelity of the WGA
in our experiment. SNPs on repeats were excluded. We observed the opposite allele calls on
only 0.73% of these homozygous SNPs. These errors may be caused by WGA errors,
Illumina genotyping errors, and overlapping with un-annotated repeats, CNVs, or segmental
duplications.
NMETH-BC07899D
Supplementary Table 4. Call rates of the Illumina genotyping with whole genome
amplification products of microdissection harvests in the 7DDNA haplotyping.
.
Haploid chromosomes
Diploid chromosomes
Samples
Total SNPs
Called SNPs
%
Total SNPs
Called SNPs
%
Sample-1
148565
36961
24.9
191502
55092
28.8
Sample-2
139103
43295
31.1
200964
108057
53.8
Sample-3
67703
25232
37.3
283296
159662
56.4
All
355371
105488
29.7
675762
322811
47.8
All SNPs, including both homozygous SNPs and heterozygous SNPs reported in the
genotyping with the genomic DNA of GM10847, were included in this analysis.
NMETH-BC07899D
Supplementary Table 5. Allele dropout of the Illumina genotyping with whole genome
amplification products of microdissection harvests in the 7DDNA haplotyping.
Total
Number of
Samples
Observed
heterozygous heterozygous
chromosomes
SNPs
Observed
Obs/Total
Heterozygous
(%)
SNP/chromosome
SNPs
Haploid chromosomes
Sample-1
9
13641
8
0.059
0.9
Sample-2
10
15649
13
0.083
1.3
Sample-3
4
7193
8
0.111
2.0
All
23
36483
29
0.079
1.3
Diploid chromosomes
Sample-1
13
13241
832
6.3
64
Sample-2
12
27011
2827
10.5
236
Sample-3
18
40287
5029
12.5
279
All
43
80539
8688
10.8
202
Only heterozygous SNPs reported in the genotyping with the genomic DNA of GM10847
were included in this analysis. Those SNPs located on repeats were excluded. Repeated
were identified by RepeatMasker.
NMETH-BC07899D
Supplementary Table 6. Probability for haplotyping whole-genome after m cuts.
The number
pm ([1-P(A)])
of cuts (m)
Assumption A1
Assumption A3
1
1.02E-06
1.19E-07
2
0.00201
0.00134
3
0.05830
0.04636
4
0.25960
0.22664
5
0.52190
0.48180
6
0.72910
0.69613
7
0.85730
0.83494
8
0.92760
0.91391
9
0.96390
0.95603
10
0.98220
0.97778
11
0.99130
0.98883
12
0.99570
0.99440
13
0.99786
0.99720
14
0.99893
0.99860
15
0.99945
0.99930
16
0.99971
0.99965
17
0.99983
0.99982
18
0.99989
0.99991
19
0.99992
0.99996
20
0.99994
0.99998
The detailed calculation procedure of this table is provided in Supplementary Equation-1.
NMETH-BC07899D
Supplementary Table 7. The SNP coverage in cumulative 7DDNA haplotyping.
SNP coverage%
Number of
Haplotyped chromosomes
chromosomes Mean
SD
Min
Max
chromosomes haplotyped by 1 harvests
8
27.5
10.0 15.9
39.0
chromosomes haplotyped by 2 harvests
5
49.4
2.6
45.2
51.6
chromosomes haplotyped by 3 harvests
1
59.5
59.5
59.5
NMETH-BC07899D
Supplementary Note-1. The probability that at least one chromosome has not been
separated in cumulative haplotyping.
Assumptions:
A1: Each cut results half of chromosomes in one side and the other half in the other side;
A2: The cut from one cell is independent to the cut from another cell;
Notations

n : the number of chromosome pairs ( n  23 );

m : the number of cells used (the number of cuts);

Ai : the event that the i th chromosome pair is not separated ( i  1, , n ) after m cuts;

Aij : the event that the i th chromosome pair is not separated in the j th cut
( i  1, , n; j  1, , m )

A : the event that at least one chromosome is not separated after m cuts;
n
m
We have A   i 1 Ai and Ai   j 1 Aij . So
n
P ( A)  P( i 1 Ai )
k 
n
  k 1 (1) k 1   P( A1  A2  Ak )
n
m
k 
n
  k 1 (1) k 1   P( j 1 ( A1 j  A2 j  Akj ))
n
k 
n
  k 1 (1) k 1   [ P( A11  A21  Ak1 )]m ,
n
where A11  A21  Ak 1 is the event that the first k chromosome pairs are not separated in
one cut. We also have
NMETH-BC07899D
 l   n  2l 
 

k 2n  2k 
k
.
P ( A11  A21  Ak1 )   l 0,2l  n   
n
 
 2n 
We can understand this calculation as follows:
n
  : The total number of combinations;
 2n 
l 
 :
k 
The number of combinations that l chromosome pairs are not separated in one
side ( l chromosome pairs in one side while k  l haplotype pairs in the other side;
 n  2l 

 : The number of combinations when l chromosome pairs are not separated in
 2n  2 k 
one side;
If we change the aforementioned assumption A1 as the following assumption:
A3: each chromosome has a probability of 0.5 to enter one side in the cut,
then
n
n
P ( A)  P( i 1 Ai )  1  P(( i 1 Ai )c )
n
 1  P( i 1 Aic )  1  P( A1c ) n ,
m
m
where P ( A1c )  P (( j 1 A1 j )c )  1  P ( j 1 A1 j )  1  P ( A11 ) m  1  0.5m. So we have
P ( A)  1  (1  0.5m ) n .
Supplementary Table 6 showed the probability data that all chromosome pairs are separated
after m cuts, which is 1  P( A) .
NMETH-BC07899D
Supplementary Note-2. The estimation of 7DDNA Cost.
Assume that for each individual, we will make up to 12 microdissection sampling cut. Assume
that in each round we will process only one of an individual’s microdissection harvests (whole
genome amplification and Illumina SNP genotyping). At the end of each round, successfully
haplotyped chromosomes are summarized for each individual. When an individual has
already received a haplotype report for all 23 chromosomes, his/her remaining
microdissection harvests will not be processed in the next round. Our statistical analysis
(Supplementary Equation-1) showed that 4-8 microdissection samplings of each individual
will provide a 93% probability to obtain his/her whole genome haplotypes (Supplementary
Table 6 and Supplementary Fig. 4). In other words, ~93% of individuals will need no more
than 8 microdissections to complete his/her genome haplotyping. When we reach 12
microdissection samplings, 99.6% of all individuals will have his/her whole genome
haplotypes (Supplementary Table 6).
Assume that the cost for first round haplotyping is x and the cost for each addition round for
m repeats (from m cells) is ym (m  1, 2,) . Denote the probability of that the procedure is
repeated m times is pm (m  1, 2,) , then the expected cost of the experiment is:
x   m 1 pm ym
The probability pm is the probability that at least one chromosome pair that is not separated
after m  1 microdissections but all chromosome pairs are separated after m cuts. The pm
values are provided in the Supplementary Table 6.
NMETH-BC07899D
For each specimen, there is a basal cost on a genotyping assay with its genomic DNA, and a
cost ($5.20) on tubes, tips, chemicals, and a foiled slide. Based on the outsource service
price, $473 for Illumina Omni1-Quad BeadChip and $342 for Illumina CNV370 BeadChip,
x = $473+5.20 = $478.20 (Illumina Omni1-Quad BeadChip)
or x = $342+5.20 = $347.20 (Illumina CNV370 BeadChip)
The cost for each cut includes a collecting tube ($0.07), whole genome amplification (WGA)
($10) and a high-throughput SNP genotyping ($473).
So
ym= $483 (Illumina Omni1-Quad BeadChip) or $352 (Illumina CNV370 BeadChip)
ym =
ymx m
The total outsource service cost per individual will be,
(Illumina Omni1-Quad)
x   m 1 pm ym = $478 + m-1 pm x 483 = $2675/individual
(Illumina CNV370)
x   m 1 pm ym = $347 + m-1 pm x 352 = $1968/individual
* The pm values are provided in the Supplementary Table 6.
The total outsource service cost per individual per SNP will be,
(Illumina Omni1-Quad)
$2675/1140419 = $0.0023/individual/SNP
(Illumina CNV370)
$1968/351507 = $0.0056/individual/SNP