RINet
LIVE STRUCTURE / 1CRNCα CONTACT GRAPH
EXAMPLE RESULTMOST IMPORTANT RESIDUES RANKEDOPEN THE DEMO FOR MUTATIONS, COMPARISONS, AND NEXT STEPS
SUPPLIED COORDINATESLOCAL ANALYSIS · NO ACCOUNT
Protein structure analysis
Analyze a protein
structure.
Load a PDB or mmCIF file. Choose the functional site you want to test. RINet identifies the most useful residue to mutate, a comparable mutation expected to have less effect, the measurements to run, and the next construct to make.
Files are parsed on this device and never uploaded to RINet. Public PDB lookup contacts RCSB PDB.
Analysis result
Pending
REPRODUCIBLE RECEIPTPending
3D STRUCTURESupplied coordinates
CHAIN COLORSBuilding legend…
STRUCTURE MAPBUILDING COLOR-CODED SURFACE
RESIDUE TO TEST FIRSTPendingCALCULATING STRUCTURAL CONTACTS
Pending coordinate coveragePending non-local contacts
SUPPLIED COORDINATES · STATIC STRUCTURE · NOT A TRAJECTORY
Analysis ready · coordinates remained on this device.
ResiduesPending
AtomsPending
ChainsPending
Structural contactsPending
Confidence / B fieldPending
From protein structure to a lab-ready test
Find the most useful residue to mutate.
Pick the functional site. RINet ranks candidate mutations, finds a structurally matched comparison, and turns the pair into a measurable experiment.
QUESTION TO TEST
Choose the kind of residue you want to test. RINet will rerank the protein for that question.
01 / FUNCTIONAL SITE
Which part of the protein do you want to affect?
Choose a pocket, interface, or known functional residue set. The ranking updates for that site.
Building target choices from the coordinate record…
Why this residue was selected
See the score and its contact neighborhood.
The bars show why it ranks here. The map shows its nearest Cα contacts. Choose any ranked residue to update both.
SCORE BREAKDOWNPending
Select a ranked residue to inspect every score component.
MUTATION CHEMISTRY
Pending
WHY THIS COMPARISON WAS CHOSEN
Pending
CONTACT NEIGHBORHOODClick any node to select it
selected residuedirect contactother ranked residuenearby contact
Lines are Cα contacts under the active cutoff. The selected residue is fixed at the center; only its nearby contact neighborhood is shown. This is a graph layout, not a physical trajectory.
WHAT TO DO WITH THIS RESIDUE
Pending—
Structure-derived ranking; not a probability of function.
WHY THIS RESIDUE RANKS FIRST
Pending
BUILD THESE CONSTRUCTS
Pending
BUILD THIS COMPARISON
Pending
MEASURE THESE OUTPUTS
Pending
THIS RESULT SUPPORTS THE RESIDUE IF
Pending
Select a ranked residue to update the construct and measurement plan.
01 / QUESTION
Define the functional readout
Name the measurement that represents function in your system. This changes the experiment sheet, not the structural calculation.
Loading the suggested assay…
See the physical model, equation, parameters, and exact ranking
SELECTED-SITE MECHANICAL RESPONSE
Calculating the selected site response.
(H + λI)u = fₜ; Sᵢ(T) = Σe∋i ⟨[nₑ·(uₚ−u_q)]²⟩x,y,z; ΔJₜ ≈ εSᵢH is the anisotropic-network Hessian; balanced unit forces are applied to the selected site along x, y and z.
Loading exact ranking equation…02–03 / MUTATION + COMPARISON
Build the top mutation and its comparison in the same batch.
Run the proposed mutation, its matched lower-effect comparison, and wild type together. If candidate and comparison behave alike, the result does not support a site-specific effect.
BUILT-IN DEMOSynthetic results are already filled.
The % WT values below are not published 4HHB measurements. They are included only to demonstrate result classification and next-experiment selection.
| Construct | Why it is included | Residue / mutation | Structural rationale | Function % WT | Abundance % WT | Fold / assembly % WT | Interpretation |
|---|---|---|---|---|---|---|---|
| Building the first round… | |||||||
04 / MEASUREMENTSEnter three values for every construct.
Function answers the biological question. Abundance and fold/assembly show whether the construct still produced usable protein. Without both quality checks, a functional loss cannot be assigned to the selected site.
Construct list ready.Enter function, abundance, and fold/assembly for each construct, or use the CSV template.
05–06 / RESULT + NEXT STEP
Did the residue affect function, or did the mutation damage the protein?
Complete all three measurements for at least two proposed mutations and one comparison mutation.
No measurements entered
Enter the measurements above. RINet will then state which construct to build next and why.
How the residue ranking and comparison mutation are calculated
Every proposed mutation is ranked for the same user-chosen functional site. Its comparison residue has similar local chemistry and structure but a lower predicted effect on that site. Run both mutations to test whether the higher-ranked residue really gives a more specific result.
Use a ligand pocket, biological interface, or experimentally established residue set.
Apply balanced forces to the target and solve the anisotropic elastic network along three axes.
Repeat at 8.0, 8.5 and 9.0 Å so a fragile network cutoff does not silently drive the result.
Choose a substitution that removes or alters side-chain contacts without automatically choosing the most destructive mutation.
Test a structurally similar mutation predicted to affect the chosen site less. If both mutations behave alike, the ranking did not isolate a specific residue effect.
SCIENTIFIC BASISEstablished mechanics; a new experiment-design integration
A Cα elastic network provides the directional Hessian used for near-native collective response.
Linear response to residue forces has been used to connect structural perturbations with observed conformational change.
The 8.5 Å primary cutoff and the three-cutoff check follow published evaluation of elastic-network parameterization.
Normalizing a site's response by background compliance is established; RINet extends the idea to a selected target-region cross-compliance contrast.
Energy-weighted bond-propensity methods show why predicted long-range effects should be benchmarked against known allosteric systems.
RINet links the calculation to the lab decision: choose a functional site; calculate which residues should affect it; check whether the ranking survives three network cutoffs; choose a mutation from observed local contacts; find a structurally similar comparison mutation; collect function, abundance and fold/assembly; then choose the next construct from the result. Prospective performance must still be established on held-out systems.
COORDINATE EVIDENCE
What atoms touch this residue?
Select a residue to inspect its atom-level partners before choosing a mutation.
These atom contacts show what the side-chain replacement can change locally. The elastic-network calculation above estimates how strongly that residue is connected to the chosen functional site.
OBSERVED IN THIS MODELPending
No residue selected.
FIRST MUTATION: CONTACT CAPABILITIESPending
The mutation-specific comparison will appear here.
Side-chain replacement changes atom identity and geometry. A modeled mutant, energetic calculation, and experiment are still required.
WHY THIS RESIDUE RANKS ABOVE ITS COMPARISONPending
The exact score differences will appear here.
Method scope
What RINet calculates
Open the four-part method summary +
Method scope
What RINet calculates
Author chain IDs, residue numbering, ligands, missing atoms and quality fields remain visible.
Directional elastic response, normalized cross-compliance, cutoff robustness, and mutation-specific local constraints.
Construct list, function and protein-quality measurements, result interpretation, and the next construct to build.
Use atomistic energy, ensembles or MD as a separate layer when relaxation, solvent, kinetics or large state changes determine the question.
Sequence and coordinates
Sequence mapped to author residue numbering
Author chain IDs and residue numbers are preserved. Missing coordinates and insertion codes are shown.
primaryrankedselectedCDR heuristicConservation is not scored until an external sequence analysis is supplied.
Mechanical-model checks
Does the target response survive basic challenges?
Check known sites, cutoff sensitivity, another conformation, or your own mutational labels.
Checking benchmark manifest…
Target-sensitivity stability at 8.0, 8.5 and 9.0 Å
Large sensitivity-rank changes indicate dependence on network construction and lower the displayed robustness term.
Compare the same protein in another state
No second state loaded. RINet will compare matched chain/residue identifiers, contact changes and rank shifts.
Reference check, not general validation
The built-in known sites test whether the ranking recovers established structural anchors in selected examples. It is not a substitute for blinded multi-protein benchmarking against mutational, allosteric, evolutionary or MD datasets.
Test this ranking against your own mutational labels
Import positive and negative residues for this structure. RINet reports AUROC for the complete score, target sensitivity, target distance and Cα degree. Optional conservation, MD or external scores are evaluated beside them.
No label set imported. Required columns: chain, residue, known_functional. Optional: insertion, conservation_score, md_score, external_score.
Experiment for the selected residue
Build these constructs, run these measurements, then make this decision
CALCULATION DETAILSREPRODUCIBLE RANKINGThe score equation and structural evidence are shown above.
The structural question will appear here.
The mutations and measurements will appear here.
The result to look for will appear here.
The next action will appear here.
Experiment template
Select the intended use
Choose the kind of study you are running. This updates the measurements below; it does not change the residue ranking.
BIOLOGY MODE
General structure–function test
Build the highest-ranked mutation and its structurally matched comparison.
Source: supplied PDB metadata, coordinates, and residue graph.
Build the selected mutation, its comparison mutation, and wild type.
Measure molecular integrity before the protein-specific functional readout.
Did the proposed mutation change function more than the comparison mutation while the fold remained intact?
WHAT CHANGES
The study type changes the measurement plan. The selected residue and comparison mutation still come from the structure calculation above.
EXPERIMENT PLAN
Selected-residue assay plan
Begin with the smallest chemical perturbation. Measure protein quality before interpreting the functional readout.
Side-chain perturbations
- First probePendingSmallest interpretable property change available; not assumed to be functionally conservative.
- Property removalPendingRemoves or reduces the main side-chain property.
- Stronger follow-upPendingUsed only after the first construct can be interpreted.
Measurements
- 0
Confirm author numbering, expression-construct sequence, and assay dynamic range.
- 1
Check abundance plus one folding or stability readout.
- 2
Measure the protein-specific functional output.
- 3
Run the proposed mutation, comparison mutation, and wild type in the same batch.
Four common outcomes
Function changes; quality holds
Repeat the result, then test a stronger mutation at the same residue.
Function and quality both fall
Treat as a stability, expression, or assembly defect first.
Both mutations behave alike
Do not prioritize this residue; test the next-ranked residue or revise the assay.
No measurable effect
Check assay sensitivity, then build the next-ranked mutation.
What else could explain the result?
- Local packing change rather than a specific functional pathway.
- Expression, trafficking or folding loss rather than site-specific biology.
- The supplied static state may not represent the assay-relevant conformation.
IF INCONCLUSIVE, BUILD THIS NEXTPending
Interpretation limitThe structure does not specify assay conditions, cellular context, replicate count, or a meaningful effect-size threshold. Set these from the experimental system before ordering constructs.
Scientific evidence and methodsScore equations · atomic contacts · prior art · full ranking · reproducible methods
02
Coordinate record
| Chain | Residues | Atoms | Sequence span | Gaps / breaks |
|---|
03
Structural flags
04
Contact-dense residues
Cα ≤ 8.0 Å · adjacent residues excluded
Contact degree is a geometric descriptor, not proof of functional importance or allosteric causality.
05
Complete expert ranking
All parsed residues · sortable evidence record
| Rank | Residue | Score | Site-effect %ile | Ranking stability | Site distance Å | Contacts changed | Atomic partners | Polar | Ionic | Cα degree | Betweenness | Interchain | Why |
|---|
06
Methods text
07
Interpretation limits
Read before ordering constructs
The elastic response is a linear, near-native-state calculation from one conformation. Compare relevant states or use MD when anharmonic motion, kinetics, solvent, membranes, or large rearrangements matter.
The anisotropic network estimates target-to-residue compliance. Mutation relaxation, atomistic energetics and binding or folding free energy require a separate validated calculation or experiment.
The supplied coordinate file may contain an asymmetric unit, crystal contacts or an incomplete biological assembly.
Author numbering, insertion codes, unresolved residues, tags and assay constructs must be reconciled experimentally.
Conservation is not invented. Use the FASTA/BLAST route or import external evidence before making evolutionary claims.
Known-site recovery and cutoff stability support inspection; they do not establish general predictive performance.
Export
Save the analysis
Download a local JSON record of the structure summary, ranked targets and reproducibility details. Nothing is published.