The Opportunity
Offload execution to Fabric. Focus on what makes BioBuild unique: the AI that understands scientific questions.
BioBuild's Moat
Knowing when to use DESeq2 vs edgeR, Cox vs Fine-Gray. This expertise stays with BioBuild.
Clean Separation
BioBuild: AI + UX + Scientific Knowledge
Fabric: Execution Engine + Workflows
BIOBUILD
SCIENTIFIC INTENT PARSER
"Find genes causing drug resistance in my cohort"
▼
KNOWLEDGE-DRIVEN METHOD SELECTOR
"Use DESeq2 for low counts, Cox for survival data"
BioBuild's Moat
▼
PIPELINE SYNTHESIZER
Step 1
→
Step 2
→
Step 3
→
Step 4
▼
▼
▼
▼
EXECUTION ENGINE
powered by Fabric: 21 Pre-built Bioinformatics Workflows
Sequence
GC, Align, Translate
Gene Analysis
Codon, ORF, Tm
Protein
Domains, Structure
Quality
Filter, QC, Stats
Simple API: workflow: "reverse_complement" then results in under 1 second
[ Distributed Agent Network / Auto-scaling / 100% Reliable ]
▼
▼
▼
▼
RESULTS + PROVENANCE + VISUALIZATION
Proof of Concept: Live Demo Results
21
Workflows
Pre-built bioinformatics operations ready to call
<1s
Execution Time
Per workflow call
100%
Reliability
No syntax errors, no retries, production tested
Integration model: BioBuild's AI understands "calculate reverse complement" then calls workflow: "reverse_complement" then Fabric executes instantly then results returned. No code generation needed.
Workflow API
One parameter: workflow: "codon_analysis". Fabric handles everything else. No code generation needed.
AI Integration
BioBuild's AI maps user intent to a workflow name. Fabric executes it in under 1 second. Clean separation of concerns.
21 Workflows
Sequence analysis, alignment, codon optimization, translation, protein domains. All ready.
Execution Speed
<1 sec
per workflow call
Reliability
100%
no syntax errors, no retries
Integration Path
How BioBuild would integrate Fabric workflows:
Step 1
User asks: "Calculate the reverse complement of my sequences"
BioBuild's AI understands the scientific intent
Step 2
AI maps intent to Fabric workflow: "reverse_complement"
Simple lookup table or LLM mapping. BioBuild's Method Selector handles this.
Step 3
API call: { workflow: "reverse_complement", data_url: "file.fasta" }
One HTTP POST to Fabric. No code generation, no Python execution.
Step 4
Results returned in under 1 second. BioBuild displays to user.
Fabric handles compute. BioBuild handles UX and scientific interpretation.
Example API Call
{ "workflow": "reverse_complement", "data_url": "sequences.fasta" }
All Bioinformatics Workflows
Sequence Analysis
gc_content
sequence_features
sequence_stats
reverse_complement
Alignment
sequence_alignment
phylogenetic_distance
pattern_search
Gene Analysis
codon_analysis
sequence_translation
melting_temp
gc_skew
Protein Analysis
domain_analysis
structure_prediction_prep
sequence_features (protein)
Quality Control
quality_control
sequence_filtering
fastq_processing
Expression Prep
expression_prep
normalization
batch correction
Each workflow: workflow: "codon_analysis" + data_url: "file.fasta" = Results
Live Workflow Results Click to view screenshots

Each result produced by a single workflow: "..." API call. No code generation.

Workflow Results 1 Workflow Results 2 Workflow Results 3 Workflow Results 4 Workflow Results 5 Workflow Results 6 Workflow Results 7 Workflow Results 8 Workflow Results 9