1

1. Bionano Sample to Answer Workflow on the Saphry System 2. Bionano Prep SP Tissue & Tumor DNA Isolation

Rapid Isolation of high quality ultra-high molecular weight (UHMW) genomic DNA (gDNA) from blood, bone marrow aspirates (BMAs) and fresh frozen human tumors

H. Sadowski1, Y. Zhang1, K. Pham1, H. Way1, Jill Lai1, Andy Pang1, Alex Hastie1 and M. Oldakowski11Bionano Genomics, San Diego, CA, USA

PresenterPresentation NotesBionano Sample to Answer Workflow on the Saphry SystemProcess begins with the isolation of UHMW genomic DNA from various sample types. We’ve replaced the original plug lysis protocol with the much more rapid, more stream-lined solution phase “SP” protocol that is not only cheaper but amenable to automation. The first Bionano Prep SP kit tackled fresh/frozen cells and EDTA bloods (Population Studies and Rare Genetic Disease analysis), and the second SP kit extended to frozen bone marrow aspirates (Leukemias). After isolation, long DNA molecules are labeled by incorporation of fluorophores at a specific sequence motif throughout the genome using DLS technology. The labeled genomic DNA is then loaded into the Saphyr chip and linearized in the Nanochannel arrays. Single DNA molecules are imaged then digitized. Molecules are uniquely identifiable by the distinct distribution of sequence motif labels and then assembled by pairwise alignment into genome maps for de novo whole genome assembly (reference blind) or in comparison to a reference genome or multiple samples in the rare variant pipeline. RVP is typically used for detection of structural variation in cancers and other samples where there is potential for low allelic fraction (mosaicism).

Bionano Prep SP Tissue & Tumor DNA Isolation Protocol WorkflowWe have now developed an SP protocol for small amounts of tissues and human tumors. The Bionano Prep SP Tissue and Tumor DNA Isolation Protocol provides UHMW genomic DNA in less than 6 hours from a batch of up to 8 samples with approximately 10 mg fresh or freshly frozen tissue or tumor. It utilizes a homogenization, lyse, bind, wash, and elute procedure that is common for silica-based gDNA extraction technologies in combination with a novel paramagnetic disk. Unlike magnetic beads and silica spin columns, which shear large genomic DNA, the Nanobind Disk binds and releases genomic DNA with significantly less fragmentation. After homogenization in a buffer containing ethanol, tissue lysis and Proteinase K digestion occurs in a chaotropic buffer. The released genomic DNA binds to the Nanobind Disk upon the addition of salting buffer and isopropanol. After four wash steps, the disk is transferred to a fresh tube and the genomic DNA is eluted from the disk in 65 ul of Elution buffer. The recovered UHMW genomic DNA is subjected to limited shearing, mixing and equilibration overnight at room temperature to facilitate DNA homogeneity before the DNA concentration is determined. Typical UHMW gDNA size ranges from 50 Kbp to ≥ 1 Mbp. This protocol was developed and validated using freshly frozen tissues from Brown Norway rats as well as several human solid tumors obtained from BioIVT. The results with the human tumor samples is described in the next slide.

2

3. Characteristics of the Freshly Frozen Human Tumors from BioIVT

Tissue Amount(mg) Clinical DiagnosisTumor Grade Stage STMR SNML SNCR SOTR RIN

Date of Excision

Age at Excision Sex

Bladder 4882 Urothelial Carcinoma High Grade II 89 0 8 3 9.27 2016 64 Male

Brain 960 Astrocytoma Anaplastic III III 100 0 0 0 9.52 2013 39 Female

Breast 2886 Ductal Carcinoma In Situ High Grade IIB 97 0 1 2 10.0 2013 49 Female

Colon 2620 Adenoma Invasive Well Differentiated UNK 97 3 0 0 9.71 2015 78 Male

Kidney 3285 Renal Cell Carcinoma II II 100 0 0 0 9.71 2012 61 Female

Liver 2806 Hepatocellular Carcinoma Mod Differentiated I 88 0 0 12 9.73 2016 71 Male

Lung 4440 Adenosquamous Carcinoma Moderate to Poor IIA 100 0 0 0 9.72 2012 36 Male

Ovary 3587 SerousCarcinoma NR IA 100 0 0 0 9.54 2013 39 Female

Prostate 372 Adenocarcinoma Invasive 4+4=8 IV 92 0 0 8 9.07 2016 61 Male

Thyroid 2860 Papillary Carcinoma Well Differentiated I 100 0 0 0 9.57 2018 36 Female

4. Average Tumor Single Molecule Quality Report (MQR) Metrics

Tumor (#Samples-#Operators) Input (mg) DNA [ ] (ng/µl) DNA Yield (µg/mg) N50 Kbp(>20Kbp)N50 Kbp

(>150Kbp) Labels/100 Kbp Map Rate (%) Gbp/Scan

Bladder (3-3) 9.7 245 1.64 313 357 15.2 91.8 66

Brain (2-2) 10.5 168 1.06 228 292 14.6 90.2 42

Breast (3-2) 13.3 183 1.04 317 395 14.2 84.1 77

Colon (3-2) 11.3 231 1.33 263 330 14.9 88.6 42

Kidney (3-2) 10 96 0.63 201 269 14.6 83.5 41

Liver (2-2) 9.0 196 1.41 265 306 14.9 89.3 84

Lung (5-3) 9.6 128 0.86 248 304 15.0 90.4 51

Ovary (2-2) 10.5 168 1.05 228 292 14.6 90.2 42

Prostate (4-3) 10.3 113 0.72 273 361 14.8 85.1 62

Thyroid (4-3) 10 126 0.77 213 294 14.5 87.6 64

PresenterPresentation Notes3. Characteristics of the Freshly Frozen Human Tumors from BioIVT We obtained freshly frozen samples from 10 clinically important human solid tumor types. All of the samples chosen had been processed promptly following resection yielding high RIN#s when assayed for isolated RNA quality. The column descriptors are largely self explanatory except for:STMR The estimated %tage of the entire specimen that is included within a tumor focus. These %tages ranged from 88-100% (mostly tumor)SNML The estimated %tage of the entire specimen that is non-neoplastic or normal. These %tages ranged from 0-3% (mostly tumor)SNCR The estimated %tage of the entire specimen that is necrotic. These %tages ranged from 0-8% (not a lot of necrotic material)SOTR The estimated %tage of the entire specimen that can be described as something other than above. These %tages ranged from 0-12% (so only a little remains unassigned)

4. Average Tumor Single Molecule Quality Report (MQR) MetricsA single frozen tumor in a ziplock bag on dry ice was fragmented into small pieces by impact force. Operators picked appropriately sized fragments (samples), weighed the fragments, and each fragment was cut into smaller pieces on an aluminum block sitting on ice. The pieces from each sample were then transferred into a 15 ml tube containing 2 ml of ice cold homogenization buffer and the pieces homogenized for 10 sec at max speed with a TissueRuptor (Qiagen). UHMW gDNA was isolated as described in Figure 2. For each tumor, the number of samples analyzed and the number of different operators are given. The metric numbers in each column represent the average of the analyzed samples from each tumor. All eluted gDNAs were well above the minimal concentration required for DLS labeling (35 ng/ul) and the average final DNA yields in 65ul of elution buffer for each tumor ranged from ~6-16ug/10 mg input tissue. The MQR metrics for all of the samples easily passed specification. The filtered N50 DNA size for every sample was >250 Kbp and the average filtered N50 for each tumor ranged from ~270 to nearly 400 Kbp. The detected label densities of each sample were above 14 where the average labels/100 kbp for each tumor ranged from 14.2-15.0. Map rates for every sample exceeded 80% where the average Map rate of each tumor ranged from ~84%-91.8%. Finally the average effective DNA throughput (≥150 kbp) for each tumor ranged from 42-84 Gbp/scan.

3

5. Total Structural Variation Detected (RVP) 6. Shared Among Duo/Trio Samples (%)

7. Shared Among Duo/Trio Samples (%), Novel to Bionano control database.

*All calls are clustered. Removal of redundancy. Control filtered *Calls can be redundant. All calls clustered, No control or confidence filter

*All calls are clustered. Removal of redundancy.

Sample Deletion Insertion Duplication Inversion Interchr. translocationIntrachr.

translocation

bladder_1 317 471 57 52 1 0bladder_2 332 527 62 63 1 1bladder_3 334 503 62 52 0 0

brain_1 354 483 64 83 74 45brain_2 363 469 75 102 105 69

breast_1 314 462 57 90 58 65breast_2 335 491 80 95 61 87breast_3 326 492 73 97 61 81colon1 366 467 62 81 0 1colon3 370 481 65 83 0 1

kidney_1 281 432 51 45 0 0kidney_2 434 653 68 77 16 18kidney_3 290 443 47 43 0 0liver_1 317 486 75 71 1 4liver_3 390 580 74 69 2 5lung_1 363 492 93 96 30 49lung_2 392 514 101 95 51 55lung_3 413 556 91 101 27 35ovary_1 364 593 103 68 76 28ovary_2 375 590 90 66 70 25

prostate_1 412 584 65 86 40 28prostate_2 363 506 60 72 25 14prostate_3 327 464 53 71 24 14thyroid_1 315 452 27 49 1 0thyroid_2 350 490 55 62 1 0thyroid_3 338 478 56 64 0 1

Sample Deletion Insertion Duplication Inversion Interchr. translocation Intrachr.

translocation

bladder_1 304 (95.9) 452 (96.0) 47 (82.5) 47 (90.4) 0 (0.0) 0 (0.0)brain_1 324 (91.5) 428 (88.6) 54 (84.4) 73 (88.0) 53 (71.6) 27 (60.0)breast_1 292 (93.0) 440 (95.2) 51 (89.5) 83 (92.2) 54 (62.8) 61 (93.8)colon1 357 (97.5) 445 (95.3) 54 (87.1) 79 (97.5) 0 (0.0) 1 (100)

kidney_1 255 (90.7) 401 (92.8) 39 (76.5) 35 (77.8) 0 (0.0) 0 (0.0)liver_1 308 (97.2) 471 (96.9) 61 (81.3) 62 (87.3) 1 (100) 3 (75.0)lung_1 330 (90.9) 465 (94.5) 64 (68.8) 76 (79.2) 21 (70.0) 29 (59.2)ovary_1 346 (95.1) 533 (89.9) 84 (81.6) 63 (92.6) 69 (90.8) 25 (89.3)

prostate_1 310 (75.2) 437 (74.8) 40 (61.5) 64 (74.4) 19 (47.5) 13 (46.4)thyroid_1 302 (95.9) 425 (94.0) 22 (81.5) 43 (87.8) 0 (0.0) 0 (0.0)

Sample Deletion Insertion Duplication Inversion Interchr. translocation Intrachr.

translocation

bladder_1 31 (96.9) 27 (87.1) 3 (60.0) 4 (100) 0 (0.0) 0 (0.0)brain_1 39 (86.7) 18 (72.0) 14 (73.7) 8 (33.3) 53 (71.6) 27 (60.0)breast_1 39 (84.8) 31 (93.9) 16 (88.9) 42 (91.3) 54 (93.1) 61 (93.8)colon1 56 (96.6) 37 (82.2) 10 (90.9) 3 (50.0) 0 (0.0) 1 (100)

kidney_1 20 (74.1) 23 (82.1) 4 (40.0) 2 (50.0) 0 (0.0) 0 (0.0)liver_1 40 (87.0) 42 (89.4) 9 (52.9) 4 (28.6) 1 (100) 3 (75)lung_1 50 (69.4) 33 (78.6) 26 (53.1) 54 (62.8) 21 (70) 29 (59.2)ovary_1 82 (88.2) 140 (83.8) 40 (78.4) 4 (40.0) 69 (90.8) 25 (89.3)

prostate_1 27 (51.9) 21 (42.9) 2 (22.2) 8 (44.4) 19 (47.5) 13 (46.4)thyroid_1 21 (84.0) 22 (73.3) 2 (100) 2 (100) 0 (0.0) 0 (0.0)

PresenterPresentation Notes5. Total Structural Variation Detected (RVP)Each labeled DNA sample is run on the SaphryDC system (1.3 Tbp of DNA >150 kbp, ~400x effective coverage). Data was analyzed through the rare variant pipeline to detect structural variation by comparison of single DNA molecules and genome maps to the human reference genome GRCh38 (hg38). Calls were clustered but can be redundant. No control or confidence filtering was applied.

6. Shared Among Duo/Trio Samples (%)Calls in sample 1 (Proband) that were also observed in sample 2 (Duo) and sample 3 (Trio-when available) were tabulated. Redundant calls were removed and calls were clustered, but no control filter or confidence filters were applied.

7. Shared Among Dual/Trio Samples (%), Novel to Bionano control database. Shared calls for each tumor in 6 are subjected to filtering against the Bionano “normal” control database. Any SV seen at >5% in the Bionano control database is removed. For this presentation, we focused on analysis of the breast tumor samples (Figures 8-11).

4

9. Shared Gene Amplifications in the Breast Tumor

*Redundancy Removed, Shared, Control Filtered

8. Circos Plot of Shared SVs in the Breast Tumor

Chromosome Region Example Genes Copy Number

15q21.1-15q21.2 COPS2, FGF7, MIR4712 4

15q21.2 USP8, TRMP7, SPPL2A 5

15q24.3 PEAK1, HMG20A 4

17q12-17q21.1 FBXL20, CDK12, NeuroD, ERBB2, Grb7, LRR3C 10

17q21.33 ITGA3, COL1A, ABCC3 5

17q24.2 PRKCA, CACN5, PSMD12, MIR548D2 4

PresenterPresentation Notes8. Circos Plot of the Shared SVs in the Breast Tumor Bionano software allows for visualization of the unique structural variants shared among all three samples from the human breast tumor. They are not found in the Bionano control database, and as such, may be linked to the cancer phenotype. The pink lines identify translocations. The q arms of chromosomes 15 and 17 are clear hot spots showing many examples of inter and intra- chromosomal translocations. Additionally, chromosomes 1, 6, 9 and 14 also show translocations. The next ring out is the copy number track where normal copy number is 2 where gains are blue above the line and losses are red below line. There are clear regions of significant copy number gain on both chromosomes 15 and 17 (called out in Figure 9). The next ring out marks the chromosomal location of insertions (green), deletions (orange), inversions (light blue) and duplications (lavender). The software allows to one hover and click out for details. Finally the last ring out are the chromosomal ideograms, where X and Y are given by 23 and 24.

9. Shared Gene Amplifications in the Breast TumorThis table highlights some regions of significant copy number gain in chromosomes 15 and 17, along with some interesting genes within each of these regions. Amplifcation/high expression of several of these genes have been associated or implicated in playing a role in one or more human cancers. They include FGF7, MIR4712, PEAK1, CDK12, ERBB2, Grb7, COL1A and PRKCA.

5

10. A 19.6 Kbp Deletion of CDKN2A/2B Locus (p16INK4A , p14ARF, & p15INK4B Tumor Suppressors)MTAP CDKN2A CDKN2A-AS CDKN2B-AS1 CDKN2Bchr 9 gene name

11. A 65 Kbp Amplification of the CDK12 and ErBB2 Locus (Tumor Drivers)

FBXL20 CDK12 ERBB2 GSDMA

Copy number

10

2

chr17 gene name

Chr17 hg38

mapDuplication

PresenterPresentation NotesSome examples:10. A 19.6 Kbp Deletion of CDKN2A/2B Locus (p16INK4A , p14ARF, & p15INK4B Tumor Suppressors)A genome map with extensive single molecule support identifies a 19.6kbp deletion on chromosome 9 that includes both CDKN2A and 2B genes (tumor suppressor genes)

11. A 65 Kbp Amplification of the CDK12 and ErBB2 Locus (Tumor Drivers) Copy number trace at the top highlights extensive amplification of a region on chromosome 17 that includes the CDK12 and ERBB2 genes (tumor drivers). A genome map with extensive single molecule support (not shown for space limitation) identifies a 65 kbp duplication of this locus.

Conclusions:-The Bionano Prep SP Tissue and Tumor UHMW DNA isolation protocol facilitates structural variation analysis of a variety of solid human tumor types on the Bionano Saphyr system. -The Saphyr system can be used to accurately detect genetic mutation hallmarks in samples with cancer. These include large rearrangements ranging from translocations, within chromosome fusions, to copy number alterations. -Researchers can perform experiments to uncover somatic variation by comparing with Bionano control sample database, or against a matched pair sample. -Furthermore, Bionano SV pipelines can detect SVs with complex breakpoint structures that are recalcitrant to detection by other technologies. -Our results indicate that the Saphyr system can capture a broad spectrum of variation with functional importance, and can provide easy solutions for cancer studies.