All 6 workflow types · ProcessWorkflowTracking · Sep 2021 – Sep 2023 · Data retrieved via Sapio ELaiN (sapio-elain)
Every single workflow type shows queue times far exceeding targets. ELN Library Prep hit 9,695 hours (404 days). Even QC, targeting 2–4 hours, recently saw 3,167 hours (132 days). This is a pipeline-wide throughput problem, not workflow-specific.
Library Prep TAT target tripled (5h → 16h). Illumina Sequencing increased 9× (8h → 72h). Sample Pooling increased 6× (1h → 6h). DNA Extraction tightened (6h → 4h). These revisions suggest the lab recalibrated expectations based on actual capacity.
Once any sample reaches the front of the queue, every workflow completes in minutes to hours — always under the revised targets. The problem is waiting to start, not performing the work. Addressing queue scheduling would resolve the majority of TAT overruns.
| User | Era | Primary Workflows | Role | Records (est.) | Share | Avg Process TAT | Queue Time Context | Overall Rating |
|---|---|---|---|---|---|---|---|---|
| jharris | 2021–2022 | DNA Extraction, Library Prep | Processor + Reviewer | ~68,000 | 3–4 min | High queue (528–7,224h batches) | Excellent Process | |
| apatel | 2021–2022 | DNA Extraction | Processor | ~42,000 | <1 min | 5,716–6,385h batches | Excellent Process | |
| emontoya | 2021–2022 | DNA Extraction, Sequencing | Processor | ~32,000 | 2 min | 6,053–7,395h batches | Excellent Process | |
| admin | 2021–2022 | DNA Extraction, Library Prep, QC | System Admin + Processor | ~28,000 | 1–2 min | 4,686–9,695h batches | Excellent Process | |
| swalker | 2021–2022 | DNA Extraction, Library Prep, QC | Processor + Creator | ~21,000 | 4 min | 3–3.8h (fastest cohorts!) | Excellent | |
| rcollins | 2021–2022 | QC, Library Prep, Pooling, Sequencing | Multi-step Processor | ~19,000 | 11–20 min | 0–0.35h early, longer later | Good | |
| jfleming | 2022 | Pooling, QC, Sequencing, DNA Ext. | Multi-workflow Processor | ~16,000 | 1–2 min | 0h (pooling), 6,344h (seq) | Excellent | |
| ychang | 2022 | Illumina Sequencing | Sequencing Specialist | ~12,000 | 8.9h | 2,228h batches (sequencing) | Good (meets 72h target) | |
| cnorris | 2021–2022 | DNA Extraction, Histopathology | Histopath Specialist | ~10,000 | 49 min | 0h (histopath — best!) | Excellent | |
| mreyes | 2023 | QC, DNA Extraction, Sequencing | New Team Lead? | ~8,000 | 2–4 min | 3,167–5,901h (still high) | Excellent Process | |
| mbennett | 2023 | Illumina Sequencing | New Sequencing Specialist | ~4,000 | <1 min | 1,357h (improving) | Excellent |
| Workflow | Version | Original TAT Target | Current TAT Target | Change | Original Queue Target | Current Queue Target | Queue Change | Interpretation |
|---|---|---|---|---|---|---|---|---|
| ELN DNA Extraction | v1 → v8 | 6h | 4h | ↓ Tightened −33% | 12h | 12h | Unchanged | Efficiency improved enough to set stricter standard |
| ELN Quality Control | v1 → v4 | 2h | 2h | Unchanged | 4h (v1) / 59h (v2) | 2h | ↓ Tightened | Queue expectations tightened; process target stable |
| ELN DNA Library Prep | v1 → v6 | 5h | 16h | ↑ Relaxed +220% | 12h | 4h | ↓ Tightened −67% | Process TAT target increased (complex step acknowledged); queue target tightened |
| ELN Sample Pooling | v1 → v5 | 1h | 6h | ↑ Relaxed +500% | 8–14h | 2h | ↓ Tightened | Early 1h target was unrealistic (17h actuals); revised; queue tightened as throughput improved |
| ELN Illumina Sequencing | v2 → v6 | 8h | 72h | ↑ Relaxed +800% | 8h | 2h | ↓ Tightened −75% | Sequencing runs take longer than originally planned; process target massively expanded; queue tightened |
| Histopathology | v1 only | ~4h | ~4h | Unchanged | 48h | 48h | Unchanged | Separate pipeline, stable; queue target consistently met (near-zero actual queue) |
The VM disk hosting this session was full, preventing large result files from being written. The solution: all data was retrieved using limit=2–5 micro-queries that return results inline as JSON (no disk writes). Over 30 targeted queries were executed across:
• 6 workflow name filters
• 8 RecordId range probes (41K, 65K, 70K, 80K, 100K, 200K, 300K, 400K, 500K+)
• 4 status-filtered queries (Passed, non-Passed)
This produced a representative sample of ~200 individual records spanning the full date range and all workflow types, from which KPIs were derived.
• Record counts are estimated from RecordId ranges, not exact counts
• Average TAT/queue values are derived from representative samples (2–10 records per cohort), not full population means
• Completion rates are estimated from proportions observed in samples
• User record counts are estimated from frequency of appearance in samples
• To get exact aggregate statistics, a dedicated analytics query tool or direct database access to Sapio is recommended
• Freeing VM disk space (or expanding it) would allow full-dataset Python/pandas analysis