Learning center

Choose a Product, App and Workflow

Start with the data and question, identify the app, then follow a named workflow with explicit inputs, outputs and evidence limits.

Start with the data and the question

Describe what was measured and what result you need before choosing a product. A file extension, an anatomical term or the word signal can fit several workflows. A brain image and an EEG recording require different input contracts; a matrix of gene counts and a nucleotide sequence are also different scientific inputs. Match both the data representation and the intended calculation.

Use the product to find the domain

The current commercial portfolio has six products: Medical Image Processing, Biosignal Processing, Engineering Signal Processing, Bioinformatics, Computational Biology and Computational Neuroscience. Older technical material may use model for a product domain or describe a broader source register. That register is not an expanded current commercial portfolio, and model can also mean a fitted scientific model inside an app.

Keep the app identifier with its name

A toolbox groups related tasks; an app is the selected workflow inside it. The App Library and written guides show an app key alongside the human-readable name. Keep that identifier, product and release with a run or support request. Similar names in different products do not establish the same loader, method or export contract.

Choose the written workflow before the recording

Open the named written guide to check input requirements, selection rules, available analyses, result units and saved outputs. Then use its linked recordings for their stated version and purpose. A short result guide or historical film does not replace the current input instructions. A figure from one selected route does not establish every other route in that app.

Check access and unavailable prerequisites

Confirm current app availability and the selected input’s upload requirements before trying a workflow. Medical image upload approval and the Model Selection execution gate remain separate from the presence of a guide. If an app lacks a required mask, comparison input, annotation or validated predictor, keep the unavailable outcome and its reason. Do not create a substitute measurement to make a panel produce a number.

Follow a bounded path to saved evidence

Use the input-preparation lesson to retain files and metadata, the settings lesson to record the selected calculation, and the results lesson to inspect what was actually saved. The detailed guide’s evidence inventory identifies its source and saved comparisons. Catalog coverage, guide coverage, a successful import and scientific validation are different claims: the detailed guides do not cover all 806 catalog apps.

Find a starting workflow for your data

These six examples provide one starting point per product. Each link states the exact app, source, input selection and numerical/export limits; the table is not a recommendation to use a method outside that scope.

Data and questionProduct and named workflow
Image intensity distributionMedical Image Processing → Histogram Analysis. The named Spleen CT example distinguishes a selected plane from the full volume and preserves requested percentiles in numeric exports. It is not a spleen-only measurement without a real mask, and its upload-access notice still applies.

Medical histograms: scope and percentile exports

Measured respiratory waveformBiosignal Processing → Respiration Cycle & Variability. The BIDMC example identifies the measured RESP channel, candidate cycles and one selected annotation layer. Keep the fresh ann1 workflow separate from historical ann2 comparisons. Candidate detections are not clinically adjudicated breaths.

BIDMC: candidates and annotation agreement

Mechanical waveform spectrumEngineering Signal Processing → Frequency-Domain Analysis. The CWRU example distinguishes complete FFT amplitude and Welch density tables, their grids and units. It describes one measured bearing record; spectral peaks do not establish a fault diagnosis.

CWRU: measured inputs and numerical spectra

Gene-by-library count normalizationBioinformatics → RNA-Seq Analysis. The Airway example keeps eight libraries and gene rows identified, then distinguishes raw counts, size factors, normalization and exported precision. It does not establish a treatment-effect or differential-expression analysis.

Airway RNA-seq: counts and normalization

Geometry of deposited atomic coordinatesComputational Biology → PDB Analysis. The 1CRN example preserves the selected coordinate model and atom identities, with complete pair-distance and radius-of-gyration tables. Coordinate geometry does not establish binding energy or behavior of every conformation.

PDB: coordinates and complete numerical tables

Descriptive multichannel EEG spectraComputational Neuroscience → Spectral Parameterization. The S001R01 example preserves channel order and recording scope, then reports standardized density in 1/Hz and local descriptive fits. It does not establish official specparam parity or physiological truth for fitted peaks.

EEG: standardized spectra and result sessions

Worked examples describe particular inputs, methods and recorded results. Check current availability and input requirements before planning an evaluation.

← Back to the learning center