1
Supplementary Figure 1. Workflow used to screen for BRCA1 and BRCA2
2
mutations in our Molecular Diagnostics Unit
3
A cascade workflow for mutational analysis of BRCA1 and BRCA2 is used. Briefly,
4
MLPA (Multiplex ligation-dependent probe amplification, MRC-Holland) is performed
5
to detect large rearrangements followed by analysis of recurrent mutations using an
6
in-house designed Sequenom assay (data not shown). If negative, Conformation
7
Sensitive Capillary Electrophoresis (CSCE) analysis and sequencing of aberrant
8
patterns is performed. CSCE is a method based on heteroduplex analysis and has
9
shown sensitivity for heterozygous variants comparable to that of Sanger
10
sequencing, with a lower cost.9 The limitation of this technique is that only
11
heterozygous DNA changes are found, although this is not a drawback when
12
searching for DNA mutations responsible for autosomal dominant syndromes.
13
Hence, for the present study only variants in heterozygosity covered by both CSCE
14
and Multiplicom kits were considered. However, Sequenom genotyping of the
15
common polymorphisms included in our previous pipeline detected 99 homozygous
16
polymorphisms, all of which were confirmed in the new NGS workflow.
17
18
Supplementary Figure 2. Coverage distribution
19
Coverage distribution for each amplicon of BRCA1 and BRCA2 in 5 MIDs of Runs
20
R5 and R7, from Experiments 2 (A) and 3 (B), respectively. The red line indicates the
21
minimum coverage threshold of 38x.
22
23
Supplementary Figure 3. NGS workflow showing true and false positives
24
resulting from each step
1
25
Proposed workflow for BRCA1 and BRCA2analysis, indicating the number of true
26
positives (TP) and false positives (FP) resulting from each technique, filtering or
27
inspection step. Results from the Training Set are marked with green labels, results
28
from the Validation Set are marked with blue labels. True positives filtered out by
29
filters 1 and 2 are recovered by the homopolymer testing and Sanger sequencing,
30
respectively, of regions indicated by the coverage report. The Sanger sequencing
31
load is decreased thanks to the bypass of visual classification of variants detected in
32
only one strand, indicated by VIP as having forward coverage or reverse coverage of
33
zero (Fcov=0 or Rcov=0). Thanks to this workflow, all variants detected in our
34
previous workflow (227 in the Training Set and 123 Validation Set) were identified by
35
the NGS workflow and only 11 FP in the Training Set and four FP in the Validation
36
Set needed Sanger sequencing to be discarded.
37
38
Supplementary Figure 4. Venn diagrams showing similarities between filters 4,
39
5 and 6
40
Venn diagrams showing the common and different false positives (A) and true
41
positives (B) that filters 4 (green circle), 5 (blue circle) and 6 (red circle) would
42
discard. Drawn with the Venn diagram generator available at the Chris Seidel web
43
page: http://www.pangloss.com/seidel/Protocols/venn.cgi.
44
45
Supplementary Figure 5. Examples of variants in homopolymers and
46
usefulness of the HP kit
47
Three scenarios in which homopolymers cause confusion and the HP assay or the
48
visual inspection of alignment are needed to correctly classify the variant. For each
49
of them, a portion (the window cannot accommodate all the reads) of the AVA
2
50
alignment is shown, followed by a relevant portion of the MAQ-S graph for the
51
analysis of the corresponding HP assay. A) BRCA2 c.956dupA, a duplication in a
52
homopolymer of 6 nucleotides, found by VIP and confirmed by the HP assay. B)
53
BRCA1 c.1961delA, a deletion in a homopolymer of 8 nucleotides, not found by VIP
54
in the correct MID (MID6, to the left) but found in some other MIDs (one of them
55
shown to the right) and correctly detected by the HP assay. C) BRCA2 c.8946delA, a
56
deletion in a homopolymer of five nucleotides, covered by the HP assay due to its
57
proximity to a homopolymer of 7 nucleotides, compared to the same region in
58
another MID, shown to the right. Note that the sample to the right without the BRCA2
59
c.8946delA presents a double peak in a different amplicon of the HP assay, due to
60
the true positive BRCA2 c.2802_2811delACAA, not in a homopolymer but also
61
covered by the kit.
62
63
Supplementary Figure 6. Detection of large genomic rearrangements (LGRs)
64
using NGS results
65
Bar plots of the dose of NGS amplicons after normalization. X-axis: NGS amplicons.
66
Y-axis: Count ratio minus 1. Fragments with normalized ratios over 1.3 are
67
highlighted in green, indicating putative duplications. Fragments with ratios below 0.7
68
are shown in red, indicating putative deletions. A) 11 plots from control samples with
69
no alterations, plots 1, 2, 3 and 7 show false positive alterations. B) 8 plots
70
representing 8 samples with LGRs: (1) Sample with a deletion of the region
71
comprising BRCA2 exons 1-24 ; (2) Sample with a deletion comprising BRCA1
72
exons 1-13 ; (3) Sample with a deletion comprising exon 20 of BRCA1; (4) Sample
73
with a deletion comprising BRCA1 exon 2; (5) Sample with a deletion comprising
3
74
BRCA2 exon 2; (6) Sample with a deletion comprising BRCA1 exon 14; (7) Sample
75
with a deletion comprising BRCA1 exon 22; (8) Sample with a duplication of the
76
region comprising BRCA1 exons 9-24.
4