Skip to content

Vanguard Biotech Systems

INITIALIZING NEXTGENAMR INTELLIGENCE LAYER_

NextGenAMR turns bacterial whole-genome sequencing, within its published scope, into interpretable, traceable antimicrobial resistance interpretation — a controlled infrastructure layer for the laboratories, hospitals and surveillance networks of the genomic era. Interpretation is species-specific; the current operational scope is Escherichia coli.

>
Genome layerResistance layerReview layer
Product status

NextGenAMR is a functional platform in controlled preview. Internal non-clinical testing is in progress and an independent validation programme is in preparation; no independent performance results are published. The product is not presented as an autonomous diagnostic or treatment-decision system.

AMRWGSE. coliTraceabilityAccess control

Controlled preview. NextGenAMR is in internal, non-clinical testing; independent validation is in preparation. The public demonstration is an illustrative technical replay of the product's E. coli output — not a real patient or clinical sample. It is provided for illustrative purposes only, shows a model-estimated probability of resistance (never a categorical S/I/R result), and does not constitute a clinical report. NextGenAMR is not presented as an autonomous diagnostic or treatment-decision system.

DEMO OUTPUT · ECOLI_DEMO_001
nextgenamr / runs / R-2741-EC
Species
Escherichia coli
Run ID
R-2741-EC
Quality
Passed
AMR panel
9 antibiotics
AMR interpretation
generate_report
AntibioticModel estimate
Amoxicillin–clavulanic acidElevated
AmpicillinElevated
CiprofloxacinElevated
CeftazidimeNo marker detected
GentamicinNo marker detected
MeropenemNo marker detected
NitrofurantoinNo marker detected
Piperacillin–tazobactamNo marker detected
Trimethoprim–sulfamethoxazoleElevated
>
>
>
>
>
Evidence summary
  • AMR genes7
  • Point mutations3
  • Plasmid replicons2
  • Mean coverage82×
Run provenance
  • pipeline · ngamr-core@0.4.2
  • db · card-2025.09 · resfinder-2025.10
  • operator · lab.ops/41
  • started · 2026-06-19 09:14 UTC
Warnings · review
  • Low-coverage region flagged for review.
  • Marker absence ≠ susceptibility; abstains out of scope — no call emitted.
Illustrative technical replay · model-estimated probability of resistance, pending independent calibration validation · not a clinical S/I/R result.Demo provenance

Illustrative technical replay — not a real patient sample and not a clinical diagnosis.

01 / System

Genomic AMR interpretation today is often spread across disconnected scripts, databases, spreadsheets and manual judgement. Workflows composed of multiple tools can make it difficult to maintain consistent versions, provenance, handoffs and reporting across teams and runs. NextGenAMR brings that work into one auditable layer.

STATE / CURRENT

Fragmented today

QC scripttaxonomy toolAMR databasespreadsheetmanual callemailed PDF
Handoffs lost between steps · no shared provenance
unify
LAYER / NEXTGENAMR

One intelligence layer

  1. 1
    Ingest isolate WGS
  2. 2
    Interpret with evidence
  3. 3
    Trace every step
  4. 4
    Emit auditable report
One traceable run · one auditable result
Integrated

One workflow instead of a fragmented toolchain spread across teams and runs.

Auditable

Every result carries its evidence, versions and provenance.

Deployable

Designed for access control in institutional environments.

02 / Pipeline

Watch a demo isolate move through the system. Each stage checks, transforms and adds evidence — from raw reads to a structured, per-antibiotic result. This is a visual simulation of the workflow, not a live analysis.

SAMPLEECOLI_DEMO_001
STAGE 07/07SIMULATION
  1. 01Quality controldone
    reads filtered · adapters trimmedreads passing QC 98.7%
  2. 02Taxonomic confirmationdone
    species confirmed within scopeEscherichia coli assignment 99.2%
  3. 03Host depletiondone
    human reads removedhost reads removed 0.4%
  4. 04Assembly & assembly QCdone
    draft genome assembledmean coverage 82×
  5. 05Gene annotationdone
    coding sequences predictedCDS predicted 4,721
  6. 06AMR detectiondone
    resistance markers matchedAMR genes · mutations 7 · 3
  7. 07Interpretation & reportrunning
    per-antibiotic model-estimated probability of resistanceantibiotic panel 9
System log
>fastp · Q30 98.7% · adapters removed
>kraken2 / bracken · Escherichia coli 99.2%
>bwa-mem2 · host reads depleted
>SPAdes · CheckM2 completeness 99.1% · 82× coverage
>pyrodigal · 4,721 CDS predicted
>AMRFinderPlus · blaTEM-1B, sul1, dfrA17 · gyrA S83L, parC S80I
>ensemble (6 models/AB) → model-estimated probability of resistance · 9-AB panel · report ready
Resistance intelligence generated
Evidence traceability completeProfessional review boundaryReport ready
Frontend simulation · illustrative demo data · not a clinical result.
03 / Graph

Select a layer, or hover a node, to follow how evidence connects — from the genome, through the markers found in it, to the antibiotics they affect, to a per-drug interpretation, up to surveillance intelligence.

The full evidence graph.

graph online
Genome
  • Isolate genomeblaTEM-1B
  • Isolate genomegyrA S83L
  • Isolate genomeparC S80I
  • Isolate genomesul1
  • Isolate genomedfrA17
Resistance
  • blaTEM-1BAmpicillin
  • gyrA S83LCiprofloxacin
  • parC S80ICiprofloxacin
  • sul1Trimethoprim–sulfamethoxazole
  • dfrA17Trimethoprim–sulfamethoxazole
Clinical
  • AmpicillinInterpretation
  • CiprofloxacinInterpretation
  • Trimethoprim–sulfamethoxazoleInterpretation
Surveillance
  • InterpretationSurveillance

Illustrative graph based on the ECOLI_DEMO_001 technical replay — not a real patient sample and not a clinical result.

Relationship model for the demo isolate ECOLI_DEMO_001:

  • Isolate genomeblaTEM-1B (genome)
  • Isolate genomegyrA S83L (genome)
  • Isolate genomeparC S80I (genome)
  • Isolate genomesul1 (genome)
  • Isolate genomedfrA17 (genome)
  • blaTEM-1BAmpicillin (resistance)
  • gyrA S83LCiprofloxacin (resistance)
  • parC S80ICiprofloxacin (resistance)
  • sul1Trimethoprim–sulfamethoxazole (resistance)
  • dfrA17Trimethoprim–sulfamethoxazole (resistance)
  • AmpicillinInterpretation (clinical)
  • CiprofloxacinInterpretation (clinical)
  • Trimethoprim–sulfamethoxazoleInterpretation (clinical)
  • InterpretationSurveillance (surveillance)
04 / Evidence

The surfaces NextGenAMR produces: interpretation, provenance, access, database integrity, evidence and the professional-review boundary. Shown here as an illustrative technical replay of the product's E. coli output — not a real patient or clinical sample, for illustrative purposes only, not a clinical report.

interpretation · CIPIllustrative technical replay
Resistance estimate: elevatedAMR
evidencegyrA S83L
parC S80I
basisdetected determinants
prudent semantics · review required
run provenanceIllustrative technical replay
  • pipelinengamr-core@0.4.2
  • dbcard-2025.09
  • resfinder-2025.10
  • operatorlab.ops/41
  • started2026-06-19 09:14 UTC
  • readsfastq.gz · paired-end
report integrity
sha256:9f2a…c71e
run provenance recorded
device trustIllustrative technical replay
Workstation LAB-07APPROVED
MFAverified
RBACmolecular-micro-lead
access controlled
database manifestIllustrative technical replay
  • CARD 2025.09 verified
  • ResFinder 2025.10 verified
  • PointFinder 2025.10 verified
database manifest verified
evidence · blaTEM-1BIllustrative technical replay
identity100.0%
coverage100.0%
contignode_12
methodAMRFinderPlus
evidence graph loaded
review boundaryIllustrative technical replay
Outputs for professional review

Supports qualified laboratory professionals under human supervision. It does not replace phenotypic AST and is not an autonomous diagnosis.

abstention example
Low-confidence samplenon-emission
The system abstains rather than interpret a low-confidence or mixed sample.
professional review boundary
DATA-DRIVEN AMR INTELLIGENCE

NextGenAMR turns bacterial whole-genome sequencing into structured, traceable resistance interpretation, within its published scope. Every isolate is identified taxonomically, and interpretation is species-specific — each supported species with its own antibiotic panel and model set. The current operational scope is Escherichia coli, with a closed 9-antibiotic panel; additional species are in internal development.

AMR is not an abstract problem. It is measurable.

A species-specific platform

Platform architecture

NextGenAMR is not built around a single organism: a species-agnostic layer identifies the isolate, and interpretation is species-specific — each supported species with its own antibiotic panel, feature space and model set. The current operational scope is Escherichia coli; additional species are in internal development, and every call is traced end to end.

01 · Identify

Taxonomic classification of supported cultured-isolate WGS inputs (kraken2 / Bracken), gated by a species-confidence threshold.

02 · Interpret per species

Species-specific antibiotic panels and a six-model ML ensemble per species–antibiotic pair, emitting model-estimated probabilities of resistance.

03 · Trace & report

Every call carries its evidence, versions and provenance in auditable, reproducible reports (JSON + PDF).

E. coli operationalSpecies-specific panels & modelsAbstains where unsupported

NextGenAMR is in internal, non-clinical validation. Outputs are model-estimated probabilities of resistance — not a categorical S/I/R result and not a clinical diagnosis.

Attributable deaths

Estimated global deaths directly attributable to bacterial antimicrobial resistance in 2019.

WHO / Global Research on Antimicrobial Resistance, 2019 estimates
Associated deaths

Estimated global deaths associated with bacterial antimicrobial resistance in 2019.

WHO / Global Research on Antimicrobial Resistance, 2019 estimates
EU/EEA annual deaths

Estimated annual deaths in the EU/EEA directly caused by antimicrobial-resistant infections.

ECDC
Estimated yearly cost

Approximate annual cost of AMR (USD PPP) across 34 OECD / EU / EEA countries, including health-system and broader economic impact.

OECD (34 OECD/EU/EEA countries)
WHO priority bacterial pathogens

Bacterial pathogens on the 2024 WHO Bacterial Priority Pathogens List — a priority list for R&D, not a list of species supported by NextGenAMR.

WHO BPPL 2024 (priority list)
3GC-resistant E. coli BSI incidence

Estimated EU/EEA incidence per 100,000 population of bloodstream infections caused by third-generation cephalosporin-resistant Escherichia coli (ECDC EARS-Net).

ECDC EARS-Net (per 100,000, EU/EEA)

Global AMR burden, 2019

Deaths directly attributable to bacterial AMR versus deaths associated with bacterial AMR.

Attributable deaths
Associated deaths
Source: WHO / Global Research on Antimicrobial Resistance estimates.

Species-specific by design

NextGenAMR treats AMR interpretation as a species-specific problem: each supported bacterial species has its own antibiotic panel, feature space, model set and interpretation logic. The current published operational scope is Escherichia coli, with a closed 9-antibiotic panel; additional species are in internal development. The architecture is species-specific by design and built to expand; every isolate is identified taxonomically, and the system abstains safely where a species is outside its published scope.

  • Species-specific antibiotic panels
  • A model set per species–antibiotic pair
  • Abstains where a species is out of scope
EU · 2024

Increase in EU incidence of third-generation cephalosporin-resistant E. coli bloodstream infections in 2024 compared with 2019.

ECDC EARS-Net 2024
EU · 2024

Increase in EU incidence of carbapenem-resistant Klebsiella pneumoniae bloodstream infections in 2024 compared with 2019. Epidemiological context only — K. pneumoniae is not within NextGenAMR's operational scope.

ECDC EARS-Net 2024

NextGenAMR platform scope

Species-specific by design: the taxonomy layer identifies the isolate, and each supported species is interpreted through its own antibiotic panel and model set. The current published operational scope is Escherichia coli; the system abstains where a species is out of scope.

Coverage model

Taxonomy across bacterial isolates; species-specific antibiotic panels, feature spaces and model sets. Current published scope: Escherichia coli; additional species in internal development.

Antibiotic panels

Each species has its own defined, closed antibiotic panel — not a single global list.

ML models per antibiotic

Ensemble — logistic regression, random forest, gradient boosting, neural net, stacking and probability calibration.

Pipeline stages

From raw sequencing reads to a structured, traceable AMR report (JSON + PDF).

Input type

Designed around bacterial whole-genome sequencing of cultured isolates.

Execution model

Built around provenance, controlled workflow execution and auditability.

Species-specific antibiotic panel — worked example

Antibiotic panels are species-specific. The panel below is shown for one species as a technical example — each species defines its own.

AntibioticStatus
Amoxicillin–clavulanic acidExample
AmpicillinExample
CiprofloxacinExample
CeftazidimeExample
GentamicinExample
MeropenemExample
NitrofurantoinExample
Piperacillin–tazobactamExample
Trimethoprim–sulfamethoxazoleExample

Validation metrics to track

Outputs are model-estimated probabilities of resistance under an internal-validation-only posture — no clinical performance metrics are published yet. Before expansion, the priority is validation, reproducibility and operational reliability.

MetricCurrent statusWhy it matters
Number of isolates testedInternal validation in progressDefines the size of the validation dataset
Species coverageE. coli operational · species-specificSpecies-specific panels, models and interpretation; more in internal development
Antibiotics evaluatedPer-species panelsEach species has its own defined antibiotic panel
Output semanticsModel-estimated probability of resistanceResistance probability, never a categorical S/I/R
Agreement with reference ASTTo be reported after validationMeasures genotype–phenotype concordance
Major errorsTo be trackedCritical for safety analysis
Very major errorsTo be trackedCritical for resistant/susceptible misclassification risk
Failed runsTo be trackedMeasures operational reliability
Median runtimeTo be measured across pilot samplesShows real-world execution performance
Provenance completenessDesigned into workflowSupports auditability and reproducibility

Public-health figures are external WHO, ECDC and OECD estimates (epidemiological context, not NextGenAMR performance), catalogued with body, year and population scope in the citation source of truth (src/domain/evidence/amr-sources.ts); sources consulted 2026-07-15. Product metrics describe current NextGenAMR platform scope and should not be interpreted as clinical validation claims.

Capabilities

Species-specific analysis

Taxonomic identification across bacterial isolates, with species-specific AMR interpretation — each species with its own antibiotic panel and model set — and safe abstention where a species is out of scope.

AMR-focused reporting

Structured outputs designed around antimicrobial resistance evidence and readable reports.

Pipeline traceability

Designed to preserve provenance, execution metadata and reproducibility signals.

Quality control

Quality, contamination and workflow checks help identify weak or invalid inputs.

Secure architecture

Built with a security-first mindset for access control, auditability and responsible deployment.

Clinical workflow awareness

Designed with laboratories, microbiology teams and hospital environments in mind.

Fast review experience

A clean interface that helps users move quickly from results to understanding.

Expandable foundation

Species-specific by design, built to evolve across organisms, panels and institutional needs as the published scope expands.

Adaptive

Who's reading?

Tune the system to your context. Your choice reshapes what the page foregrounds — nothing is hidden, only re-ordered.

Neutral view

Select a context to tune the view. Or keep the neutral, full read.

Use cases

01 / 05

Hospital laboratories

Support genomic AMR review and structured reporting workflows.

Discuss this scenario
02 / 05

Microbiology teams

Organize resistance-related evidence in a clearer, more usable format.

Discuss this scenario
03 / 05

Public health surveillance

Help transform genomic data into comparable, auditable AMR intelligence.

Discuss this scenario
04 / 05

Research groups

Accelerate exploratory AMR analysis with a focused, reproducible workflow.

Discuss this scenario
05 / 05

Pilot programs

Deploy a controlled platform for institutional evaluation and AMR innovation projects.

Discuss this scenario
05 / Trust

Every result should be explainable, every action traceable, every deployment controlled. Trust is not a badge on this page — it is how the system is built. States below are honest: what is designed in, what is an enforced boundary, and what is still on the roadmap.

NextGenAMR core
System integrity
By design
Access control
MFA · RBAC · institutional access
By design
Append-only audit trail
Every action logged · no delete path · tamper-evident
By design
Run provenance
Versions · inputs · SHA-256 integrity
Roadmap
Report signing
Qualified (eIDAS) signature — roadmap
By design
Controlled deployment
Defined environments · versioned deps
Enforced
Clinical boundaries
Outputs for professional review · abstention · not autonomous
States:By designEnforcedRoadmap

NextGenAMR is in internal, non-clinical validation. It carries no regulatory certification (CE-IVD, IVDR, ISO 13485, FDA); its technical documentation is being prepared with future IVDR requirements in mind. Audit integrity is tamper-evident — changes are detectable, not prevented — and the software does not replace phenotypic AST or a clinician's judgement.

Built from the inside out

Vanguard Biotech Systems (VBS) is a biotech software company building secure bioinformatics infrastructure for genomic antimicrobial resistance. Its conviction: AMR interpretation should not stay trapped between fragmented tools, slow workflows and processes no one can audit. NextGenAMR is VBS's first product — a species-aware, controlled layer where bioinformatics, machine-learning interpretation, security-aware design and traceability come together — and the foundation the company is building on toward broader bacterial-genomics and surveillance workflows.

We are not building another dashboard. We are building the operating layer for AMR intelligence.

Minimal by design. Rigorous by default. Systems that look simple because the complexity has been controlled.

Founder
Juan Manuel Gómez Vargas
Founder & CEO, Vanguard Biotech Systems

Juan Manuel Gómez Vargas founded Vanguard Biotech Systems to build NextGenAMR — infrastructure that turns bacterial whole-genome sequencing into traceable antimicrobial resistance interpretation. His focus: rigorous bioinformatics, security-aware design and honest, auditable interpretation over hype.

Read the founder profile
FAQ

What is NextGenAMR?

NextGenAMR is a species-specific software platform by Vanguard Biotech Systems that turns bacterial whole-genome sequencing (WGS), within its published scope, into interpretable, traceable antimicrobial resistance (AMR) interpretation — from quality control and taxonomy through species-specific AMR interpretation (each supported species with its own antibiotic panel and model set) to auditable reporting. Its current operational scope is Escherichia coli.

Who is NextGenAMR built for?

Clinical microbiology laboratories, hospitals, research institutions and public-health / AMR surveillance networks that work with bacterial genomic data and need structured, traceable resistance interpretation.

Is NextGenAMR clinically validated?

No. NextGenAMR produces outputs for professional review, not a clinical diagnosis and not a regulatory-cleared medical device. It is in active development and internal, non-clinical validation; public demonstrations are illustrative technical replays of the product's E. coli output — not real patient samples. It carries no CE/IVDR/ISO certification, does not replace phenotypic antibiogram (AST), and clinical judgement remains with qualified professionals.

Does NextGenAMR replace clinical judgement?

No. It structures and traces genomic evidence to support experts; interpretation and clinical decisions stay with the responsible clinicians and microbiologists.

What organisms are currently in scope?

NextGenAMR is species-specific by design: its taxonomy layer identifies the isolate, and AMR interpretation is species-specific — each supported species with its own antibiotic panel, feature space and model set. The current published operational scope is Escherichia coli, with a closed 9-antibiotic panel; additional species are in internal development and not yet available. Where a species is outside the published scope, the system abstains safely. Public demonstrations are illustrative technical replays of the product's output — not real patient samples — with Escherichia coli as the worked example.

How can an institution get in touch?

We open non-clinical evaluations case by case. Institutions can reach us through the contact page; every request is reviewed for institutional fit and non-clinical evaluation readiness.

06 / Access

NextGenAMR is in internal, non-clinical validation. We are opening conversations with clinical, research and surveillance institutions interested in future controlled, non-clinical evaluation. Tell us your context — every request is reviewed manually for institutional fit and evaluation readiness.

Email
vanguardbiotechsystems@gmail.com
Location
Granada, Spain
Focus
AMR · WGS · Clinical bioinformatics

We open controlled, non-clinical evaluations case by case. Tell us your context from the contact page — we do not ask for patient data, samples, sequences or results.

Acknowledgements

This project would not exist without the people behind it. To the team — Ángel, Alejandro, Juan Carlos and Irene — for the work, the care and the long hours. To José Luis, for believing in it when it was only an idea. To my family, and above all to my parents, for everything. And to everyone who, directly or indirectly, made NextGenAMR possible — thank you, always.