The Genome Detective

How AI-Powered Software Is Cracking Cancer's Code

Article Navigation

Imagine a lab where technicians juggle six different software systems, cross-referencing spreadsheets and struggling to spot patterns in a genomic haystack. For decades, this was the reality of cancer genomics. Now, a single AI-powered platform is turning chaos into clarity—and revealing cancers' deepest secrets.

The Fragmented World of Cancer Genomics

Hematological malignancies like leukemia and lymphoma are driven by structural variants (SVs)—chromosomal breakages, swaps, and duplications that traditional tests often miss. Standard diagnosis requires a battery of tests:

Karyotyping

Low-resolution chromosome snapshots

FISH

Targeted probes for known mutations

Microarrays

Copy-number scans

NGS

DNA sequencing for small mutations 9

This fragmented approach leaves labs drowning in disconnected data. Critically, 30% of aggressive blood cancers harbor cryptic SVs invisible to standard methods—until now 5 .

VIAâ„¢: The Unified Genome Interpreter

Bionano's VIAâ„¢ software acts as a "central command" for genomic data, integrating three critical technologies:

1. Optical Genome Mapping (OGM)

Images ultra-long DNA strands to detect SVs at 10,000x higher resolution than karyotyping 5 .

2. Next-Generation Sequencing (NGS)

Identifies gene-level mutations.

3. Microarrays

Flags large DNA deletions/duplications 1 7 .

Performance Comparison

Metric Traditional Methods VIA + OGM
Resolution >5 Mb events 500 bp events
SV Detection Limited to known targets All classes (balanced/unbalanced)
Workflow 3-4 separate tests Single assay
Turnaround 7-14 days 3 days

Data sources: 4 9

Decoding Blood Cancers: The Hematology Workflow

VIA's breakthrough lies in its hematology-specific AI engine, automating three critical steps:

1. Rule-Based Tiering

The software cross-references variants against global guidelines (e.g., WHO, ACMG) to auto-classify SVs:

  • Tier 1A Pathogenic variants with strong clinical evidence
  • Tier 1B Variants in cancer-associated genes 2 3
2. Complexity Scoring

VIA calculates a genome instability index based on:

  • Number of SVs >5 Mb
  • Chromothripsis (chromosomal "shattering")
  • Specific complex patterns linked to poor prognosis 3
3. Pan-Cancer Filtering

A dynamic filter highlights variants from a curated list of 1,200 cancer-related genes, slashing manual review time by 70% 2 .

Inside the Landmark Validation Study

A pivotal study validated VIA's accuracy across 60 leukemia/lymphoma cases:

Methodology
  1. Sample Prep: Ultra-high-molecular-weight DNA extracted from bone marrow/blood.
  2. OGM Processing: DNA labeled at CTTAAG sites, linearized in nanochannels, and imaged.
  3. AI Analysis: VIA's pipeline aligned data to GRCh38, comparing SVs against 179 control genomes 9 .
Results (n=60 Cases)
Sensitivity 100%
Specificity 100%
Novel SVs Found 42
Reproducibility 96.4%

Data source: 9

The Revelation: OGM + VIA detected 100% of SVs identified by traditional methods plus 42 additional pathogenic variants missed initially. One chronic lymphocytic leukemia case revealed a cryptic IGH::MYC fusion—a high-risk marker requiring aggressive therapy 9 .

The Scientist's Toolkit: Key Components in Action

Component Function
Nanobind Disks Isolate ultra-long DNA strands (>250 kb)
CTTAAG Labeling Dyes Fluorescent tags for imaging structural breaks
Stratysâ„¢ Compute GPU-accelerated server for AI analysis
Pan-Heme BED Files Curated variant filters for blood cancers
Saphyr® Chip G2.3 Nanochannel array for DNA linearization

Data sources: 9

The Future: AI That Learns Your Lab

With the 2025 VIA 7.2 upgrade, labs gain an AI that evolves with every case:

Historical Data Integration

Prior classifications train the algorithm to match institutional preferences.

Significance-Associated Phenotype (SAP) Scoring

Prioritizes variants linked to patient symptoms 4 .

Example: A lab studying myelodysplastic syndromes can configure VIA to auto-flag del(5q) variants—accelerating analysis by 50% 4 .

Conclusion: Precision Oncology's New Era

VIA isn't just software—it's a paradigm shift. By unifying fragmented data into a single, AI-driven workflow, it hands clinicians a "Google Maps" for the cancer genome.

"The potential is endless. I'm super excited to work with this in the future"

Dr. Rashmi Kanagal-Shamanna (MD Anderson) 5

For patients, this means hope: fewer missed diagnoses, faster results, and therapies matched to their cancer's unique blueprint.

To explore VIA's impact on leukemia cases, watch Dr. Sam Dougaparsad's analysis demo

bionano.com/via-software 1

References