Projects

Intro: Building Tools for Better Experimenting

Creating tools that enable open-science and improve experimental workflows.

Active development · Validation stage

6. Automated Colony Detection & Antibiotic Plate Analysis

06

A computer-vision workflow for detecting and counting bacterial colonies on antibiotic diffusion assay plates, measuring spatial regions around the antibiotic disk, and preserving scientist review as part of the analytical process.

Current stagePrototype validationDetection, review, annotation, training, and model-promotion workflows are being integrated and tested.
Primary objectiveReproducible colony quantificationReduce manual counting while retaining traceability and scientist control over ambiguous detections.
Scientific constraintHuman-in-the-loopAutomated predictions are treated as candidates until reviewed or accepted by a scientist.
01AcquirePlate image
→
02DetectCandidate colonies
→
03ReviewScientist corrections
→
04TrainCurated annotations
→
05ValidateCompare performance

Current experimental workflow

  • Detect colony candidates from uploaded plate images, including difficult small, overlapping, and low-contrast colonies.
  • Separate measurements into defined inner and outer assay regions using a consistent plate-centered spatial rule.
  • Use a 6 mm antibiotic paper disk as the physical reference for converting image-space radius measurements into millimeters.
  • Present automated detections to the scientist for correction before they become annotation or training checkpoints.
  • Save reviewed examples into selectable computer-vision profiles that can be used to train subsequent models.
  • Keep newly trained models separate from production use until explicitly promoted, preserving the distinction between an experiment and an accepted model.
  • Export counts and associated metadata even when a model has not yet been promoted.

What is being evaluated

DetectionMissed colonies vs. false positives
MorphologySmall and overlapping colonies
GeometryPlate center and region consistency
MeasurementPixel-to-mm reproducibility
ModelingTraining-set quality and model versioning
WorkflowScientist review efficiency
Scientific status: The system is currently a methods-development and validation platform. It is not presented as a finalized microbiology assay or autonomous result-producing instrument.
Field prototype · Deployed concept

5. Flow-Meter Addition to Dinoflagellate Sampler

05

Jamie Pierson's lab wanted to measure dinoflagellate density in bays and esuaries of Puerto Rico. An earlier project connected up the phytometer, and we needed to measure the flow-rate to calculate the density.

Primary objectiveQuantify sampled water volumeAdd flow-rate measurement so dinoflagellate density can be calculated from field sampling.
SystemArduino + inexpensive flow meterPortable electronics integrated into a battery-powered field setup.
EnvironmentOn-site marine samplingDesigned for portable use in bays and estuaries.
01SampleWater intake
→
02MeasureFlow rate
→
03CalculateVolume sampled
→
04EstimateCell density

System design

  • Inexpensive flow-meter.
  • Arduino based system.
  • Built into a portable "tub" for on site use with a battery.

Engineering value

MeasurementFlow-rate capture
PortabilityBattery-powered field use
IntegrationConnects physical sampling to quantitative density calculations
Project status: A practical instrumentation addition that converts a previously qualitative field workflow into one with a directly measurable sampling parameter.
Operational monitoring · Remote alerting

4. Pump Monitor for Zebra-fish Circulation

04

The ARC needs to replace filters before they are too clogged, using pressure as an excellent proxy. The system was an analog pressure-gauge that has to be checked manually. Remote monitoring allows for 24/7 access, and notification, saving time and improving responsiveness.

Primary objectiveDetect filter loading earlierUse pressure as a proxy for filter condition instead of relying on manual gauge checks.
SystemIoT pressure monitoringRemote sensing, historical visualization, and alerting.
UsersResearch aquaculture groupDesigned for the ARC at IMET.
01SenseSystem pressure
→
02TransmitIoT telemetry
→
03VisualizeDashboard + history
→
04AlertTXT + email

Current capabilities

  • IOT based, open-source.
  • Designed for the research-aquaculture group (ARC) at IMET.
  • Provides a dashboard showing the current and historical data.
  • Can send TXT and email alerts.

Operational impact

Availability24/7 remote access
ResponseAutomatic notification
MaintenancePressure used as a filter-condition proxy
Project status: The workflow replaces periodic manual observation with continuous telemetry, historical context, and alert-driven response.
Research instrumentation · Automated control

3. Six-Headed Peristaltic Pump

03

A multi-channel fluid-control system designed for biocement experiments as part of an NSF grant.

Primary objectiveAutomate repeatable fluid deliveryCoordinate multiple pumps without continuous manual intervention.
SystemSix independent pump channelsEach channel can be operated independently or as part of a programmed pattern.
Control modelIoT + browser accessRemote control and monitoring through a lab-facing interface.
01DefinePump pattern
→
02ScheduleChannel actions
→
03RunForward / reverse
→
04MonitorRemote interface

Current capabilities

  • IoT-based control and monitoring.
  • Runs pumps independently or in complex patterns, including reverse flow.
  • Open-source and remotely accessible via browser or lab interface.

Experimental value

RepeatabilityProgrammed pumping patterns
FlexibilityIndependent channel control
AccessRemote browser operation
Project status: A custom research instrument created to make complex fluid-delivery procedures programmable, repeatable, and remotely accessible.
Operational safety · Equipment monitoring

2. Freezer Alarm System

02

An alarm system for -80°C freezers intended to reduce the risk of unnoticed temperature excursions and preserve critical stored material.

Primary objectiveDetect temperature failures quicklyProvide automatic notification when freezer conditions leave the expected range.
ResilienceDesigned for outagesOperates during power outages or loss of internet connectivity.
Design goalOpen and adaptableNo subscription fee and customizable for lab-specific monitoring needs.
01SenseFreezer temperature
→
02EvaluateThreshold condition
→
03AlertText / email
→
04RespondProtect samples

Current capabilities

  • Sends text or email alerts during temperature issues.
  • Operates during power outages or internet loss.
  • Open-source, adaptable, and no subscription fee, offering a cost-effective alternative to commercial solutions.

Operational value

ResilienceContinues through outages
NotificationText and email alerting
CostNo recurring monitoring subscription
Project status: A practical monitoring system focused on reliable notification and continuity during the failure modes that matter most in cold-storage workflows.
Data acquisition · Workflow automation

1. Spectrophotometer Data Logger

01

A custom-built data logger for the Shimadzu UV 1601 spectrophotometer that captures instrument output and moves it directly into a digital analysis workflow.

Primary objectiveEliminate manual transcriptionCapture "% transmission" measurements directly from the instrument.
Data pathInstrument → cloud databaseMeasurements are centralized and immediately available for visualization and analysis.
Scientific benefitFaster time-sensitive analysisReduces friction between measurement, interpretation, and publication-ready output.
01Measure% transmission
→
02CaptureDigital logger
→
03StoreCloud database
→
04AnalyzePlots + outputs

Current capabilities

  • Captures "% transmission" values for biomass.
  • Uploads data to a cloud database, plots in real time, and allows for direct analysis or publication-ready outputs.
  • Open-source and tailored to speed up experiments, especially where time-sensitive accuracy is critical.

Workflow value

Data qualityReduces transcription steps
AccessibilityCentralized cloud records
AnalysisImmediate plotting and downstream use
Project status: A data-acquisition bridge that turns an older laboratory instrument into a connected source for realtime visualization, centralized records, and downstream analysis.