Pipeline

A source-backed workflow, in order.

The launch walkthrough uses human alpha-synuclein as the public-source worked example. Source-specific examples stay secondary to the Lineage Spatiotemporal Fragment Atlas pitch until source review supplies approved names, region summaries, and figure-ready context.

Workflow

Ordered from source intake to redacted handoff. Each step uses the same public-source worked example so the page reads as one flow rather than separate visual modules.

  1. Step 1

    Source

    Start from a public, source-grounded protein context — UniProt records, peer-reviewed annotations, and replayable sequence references. No private datasets, no unredacted intake.

    Public-source anchorHuman alpha-synuclein, 140 residues

    The live walkthrough uses public SNCA context: UniProt P37840, RCSB PDB 9A1A, and AlphaFold AF-P37840-F1. Source-specific examples remain secondary to the atlas pitch.

    UniProt P37840 · PDB 9A1A · AlphaFold AF-P37840-F1
  2. Step 2

    Sequence / region freeze

    Pin the protein sequence and the candidate region windows to a versioned snapshot so every downstream artifact is deterministically replayable.

    Residue-window map1-60 · 61-95 · 96-140

    The public example freezes three source-backed region labels: N-terminal amphipathic, NAC, and C-terminal acidic. No raw amino-acid sequence is displayed.

    UniProt P37840 region context
  3. Step 3

    Directional fragment generation

    Walk the frozen region with directional, length-bounded fragments to produce a candidate set sized for scoring while keeping individual fragment records off the public surface.

    Software output
    Overlapping fragment windowsDirectional windows over frozen context

    The figure shows window shape only: bounded overlapping regions move across public residue context while private fragment rows stay off the public surface.

    Public schematic from frozen SNCA regions
  4. Step 4

    Scoring with method history

    Score every fragment under named, version-controlled methods. The method version travels with the result so reviewers can audit which method produced which signal.

    Simulation evidence
    Scoring methodContext · aggregation · cross-seed lanes

    The public page shows scoring families and method history, not private weights, thresholds, feature vectors, or row-level values.

    Cellico Bio scoring-method label model
  5. Step 5

    Nulls and controls

    Compare scored fragments against composition-matched negative controls. Every region that ends up on the shortlist has to clear that comparison.

    Simulation evidence
    Control gateSignal must clear public-scope comparison

    Null and matched-control lanes are displayed as gate logic only. The public site does not expose control distributions or thresholds.

    Composition-aware null / matched-control gate
  6. Step 6

    Ranked shortlist

    Surface a small, top-ranked subset under declared assumptions. Rankings are computational prioritizations only — they prioritize follow-up review, not biological activity.

    Computational prioritization
    Review shortlistA/B/C review regions, no private ranks

    The launch surface can show neutral review-region labels and evidence class. It cannot show candidate identifiers, raw sequences, or private rank tables.

    Public-safe review-region abstraction
  7. Step 7

    Export

    Prepare a clean summary record with source citations, claim labels, and methodology for downstream review. The May 14 public site shows the workflow and labels only; it exposes no downloadable artifact.

    Software output
    Summary handoffManifest · citations · claim labels

    The public handoff preserves provenance, source citations, and claim ceilings while excluding raw Cellico Bio source data and scoring internals.

    Public companion export boundary
Supporting panels

Public-source context after the workflow

These panels expand the same worked example after the seven-step flow: structure context, ensemble context, redacted fragment-network projection, and biological-context atlas. They are support material, not separate unordered sections.

Structure-context visualization (public example)

Alpha-synuclein is a public example of an intrinsically disordered protein. This panel shows how public structural context and model-confidence overlays can be presented alongside Cellico Bio claim labels. It is a context visualization only, not a Cellico Bio validation result.

N-terminal amphipathic1–60NAC region61–95C-terminal acidic96–14013570105140Human α-synuclein (UniProt P37840) · 140 residuesSchematic ensemble — see source links below for primary data.Layered 2D-to-3D disorder-context view (3 conformer halos)
Public α-synuclein context shown as a layered 2D-to-3D structure/disorder schematic (Human α-synuclein (UniProt P37840), 140 residues). Relative disorder-context clouds indicate how flexible regions can be represented around public protein regions; pLDDT remains a separate model-confidence layer. Not the actual chain shape from PDB 9A1A.
  • PUBLIC_STRUCTURE_CONTEXTRCSB PDB 9A1AExperimental/integrative ensemble structure of α-synuclein monomer.
  • MODEL_CONFIDENCE_OVERLAYAlphaFold pLDDT (AF-P37840-F1)Per-residue model confidence reported by AlphaFold. Confidence is a model-quality metric — not a disorder probability.
  • PUBLIC_DISORDER_CONTEXTSchematic disorder cloudTranslucent overlay representing public-textbook intrinsic-disorder context for IDPs. Not a Cellico Bio output.
  • COMPUTATIONAL_PRIORITIZATION_ONLYCellico Bio (label only)Cellico Bio outputs in this panel are public claim labels only — no scorer internals, no candidate ranks, no row-level fragments.
RCSB PDB 9A1A
Experimental/integrative ensemble structure of α-synuclein monomer.
AlphaFold DB AF-P37840-F1
Computed model of α-synuclein.
UniProt P37840
Reference entry for human α-synuclein (gene SNCA), 140 residues.

AlphaFold pLDDT is shown only as model-confidence context and is not interpreted here as a disorder probability. This panel uses public SNCA structure/model sources for context and is not a Cellico Bio validation result.

Public ensemble context (3D schematic)

Interactive public-source ensemble context for Cellico Bio. The backbone traces and translucent clouds show a public-safe structure/disorder context view for α-synuclein; cloud intensity is a relative context cue, not a private model output. No raw sequence, row-level fragment data, candidate identifiers, scorer internals, or private values are shown.

inherited contexttissue contextchronological contextHuman α-synuclein (UniProt P37840) · 5 conformer traces (public schematic ensemble)Public-safe ensemble — see citations below for primary sources.
Interactive public-source ensemble context for Cellico Bio. The backbone traces and translucent clouds show a public-safe structure/disorder context view for α-synuclein; cloud intensity is a relative context cue, not a private model output. No raw sequence, row-level fragment data, candidate identifiers, scorer internals, or private values are shown.

α-synuclein is intrinsically disordered; this panel does not imply a single fixed 3D structure. AlphaFold pLDDT is shown only as model-confidence context, not as a disorder probability. Region-level cloud thickness is a relative public-context cue, not a measured probability or a Cellico Bio output. Full citations →

Schematic fragment-network map with X and Y axes, connected fragment nodes carrying relative disorder-context clouds, faint biological-context boundaries (inherited / tissue / chronological context), and a small subset of glinting review-priority nodes.inherited contexttissue contextchronological contextXY
Schematic fragment-network projection for Cellico Bio pipeline context. Nodes and edges are public-safe visual abstractions of fragment-neighborhood structure; each node carries a relative disorder-context cloud to indicate that fragment neighborhoods are interpreted within local biological context rather than as isolated sequence pieces. Cloud size and opacity are public-safe schematic cues, not disclosed model outputs. Highlighted regions indicate review-priority areas in the schematic. Exact scoring values, private model features, raw sequences, row-level fragment data, and candidate identifiers are withheld.

Atom-of-aging atlas

inherited contexttissue contextchronological contextDisorder-probability density cloud · public-safe schematicCellico Bio · redacted conceptual atlas
Public-safe fragment-space projection for Cellico Bio. The view abstracts how inherited, tissue, and chronological context can be layered over a disorder-probability cloud to guide review regions. It does not show exact fragment locations, raw sequences, row-level fragment data, private scoring internals, or candidate identifiers.

Redacted conceptual atlas. The cloud, shells, rail markers, and review-region glints are public-safe schematic cues — not private model outputs, not exact protein loci, not validated biology. A data-grounded version sits behind a future internal review surface, not on this public page.