ProteinIQ

Updates

Changelog

Everything new on the ProteinIQ platform.

Mol-1 and Prompteina arrive in beta for predicting biomolecular complexes and for designing proteins from a plain English description

OpenFold3 now runs OpenBind-0, and AutoDock Vina jobs can queue PoseBusters, ProLIF or OpenMM to run as soon as docking finishes. The sidebar has pinned tools and a menu for recent jobs, and more tools catch bad inputs before a run.

Added

  • Added Mol-1 (beta) for predicting protein, DNA, RNA, ligand and ion complexes, with example jobs
  • Added Prompteina (beta) for designing protein monomers and binders from a text prompt
  • Added the IC50 to Ki calculator
  • Added OpenBind-0 weights, pocket constraints, a Use templates switch and cyclic or modified chains to OpenFold3
  • Added follow-ups to AutoDock Vina: choose PoseBusters, ProLIF or OpenMM before submitting and they run on the docked poses when the job completes
  • Added pinned tools, a New job button and recent jobs with status dots and a rename, share, move and delete menu to the sidebar
  • Added a model picker to the AI Assistant
  • Added mmCIF input to Aggrescan3D, ESM-IF1, PROPKA and RoseTTAFold3

Improved

  • Improved next steps on results: ten more tools suggest what to run next, and each suggestion explains the tool and names the file it sends
  • Improved BioEmu and BioEmu2 results with a sample list where checked samples overlay the one you are viewing
  • Improved Boltz-2.1 and the other Boltz tools so they return the full Boltz result archive and predicted ligand structures
  • Improved the tools directory with sorting and a list view, and added tool search to the app home page
  • Improved loading speed of results, guides and articles, and moving between tools no longer reloads the sidebar
  • Improved input checks in AutoDock Vina, GROMACS, IQ-TREE, Clustal Omega, Codon optimization and CleaveNet so unsupported inputs get a clear message before submission
  • Improved Oligo Analyzer to accept oligos from 8 to 200 nt and added DMSO and formamide settings for Tm

Fixed

  • Fixed eToxPred Tox-scores that differed from the original model
  • Fixed Boltz tools ignoring residue modifications, cyclic chains, hydrogens in uploaded PDB files and custom MSAs in Boltz-2.1
  • Fixed GROMACS and OpenMM rejecting docked protein–DNA and protein–RNA complexes
  • Fixed SMILES validation rejecting valid hypervalent halogens, aromatic Te and Si, and %(n) ring closures
  • Fixed job renames not saving
  • Fixed RMSD calculator failing on single-chain structures with different chain IDs, and RFdiffusion3 dropping HETATM peptide residues such as AIB from targets

RFD4-Proteina arrives in beta for designing binders, enzymes and ligand pockets, and BioEmu2 samples equilibrium ensembles of proteins, nucleic acids and ligands

Results now suggest the structure checks that usually come next, tool pages load faster on phones, and many hosted tools explain bad inputs before a run or name the exact problem when one fails.

Added

  • Added RFD4-Proteina (beta) for designing protein binders, small-molecule binders, enzymes and motif scaffolds
  • Added BioEmu2 for equilibrium ensembles and time-lagged propagation, and BioEmu now takes a protein sequence and retrieves its alignment for you
  • Added next-step suggestions that pass the structure along, such as MolProbity and fpocket after structure predictions, PROPKA after PDBFixer and ProLIF after AutoDock Vina
  • Added 2D interaction diagrams to ProLIF results
  • Added PDB and mmCIF input directly in the AlphaFold2 protein field, N-terminal acetylation and C-terminal amidation to the Molecular weight calculator, and a relax added atoms setting to PDBFixer
  • Added a Run as batch link on batch-capable tools that opens a new batch with the tool already chosen, and a menu bar for the app sidebar on phones and tablets

Improved

Fixed

  • Fixed Protenix multi-seed runs returning only some seeds' structures, and Boltz-2 failing on large complexes or labelling templates with chain letters constraints could not use
  • Fixed AutoDock Vina AutoDock4Zn mode failing on zinc receptors, automatic pocket boxes splitting linked bound ligands, and PDBFixer structures that Vina then rejected
  • Fixed the SASA calculator leaving out the GFP chromophore and other modified residues, and counting covalently bound ligands as protein
  • Fixed DeepRank-Ab failing on structure names with spaces, DLKcat misreading tab-indented substrates, and MolProbity misreading AlphaFold Server mmCIF files
  • Fixed jobs being named after the input type instead of the uploaded file, tool pages briefly asking signed-in users to sign in, and forms showing required-field errors before you press Submit
  • Fixed file uploads on the Files page, structure viewer downloads, and large FASTA previews freezing the output tab

ANARCI now renumbers antibody and TCR structures, and MD trajectory analysis results read as one report with publication-ready figures

Result pages have a new header with citations, a fuller download menu and next steps, and many hosted tools now accept more valid inputs and explain the ones they cannot run.

Added

  • Added antibody and TCR structure renumbering to ANARCI, and PDF, SVG, PNG and TIFF figure export at journal column widths to MD trajectory analysis
  • Added a citation popover, next-step suggestions and a Download all menu with current view, pose and table workbook downloads to the result header
  • Added a redesigned structure viewer with tabbed details, a metric plot and per-residue pLDDT views
  • Added settings presets to AutoDock Vina, PDBFixer and Boltz-2
  • Added diffusion samples and a seed to ESMFold2, nucleotide alignment options to IgBLAST, a custom prevalence threshold to BioPhi, all method conditioning types to Boltz-2, suboptimal duplexes to RNAduplex, Ramachandran Z-scores to MolProbity, evaluation plots to CleaveNet and a per-residue table to Chou-Fasman
  • Added job inputs, settings and errors to teammates' jobs, so a team can see how each job was run

Improved

Fixed

Eight new chemistry and sequence tools arrive, along with NuCaliby for designing coding DNA from protein backbones and task-based tool search

Workspace pages, job submission and tool search are faster, and many hosted tools now catch bad inputs before a run starts.

Added

  • Added Molecular similarity search, Chemical substructure search, Butina clustering, the Stereoisomer generator, Dimorphite-DL for ligand protonation and Codon optimization
  • Added SMILES to PDBQT and PDBQT to SDF converters for docking files
  • Added NuCaliby for designing coding DNA and protein sequences from backbones, with an interactive 3D viewer for backbone ensembles
  • Added task-based tool search, so searching "protein protein docking" or "folding" finds the right tools, plus a Task filter on the tools page
  • Added a downloadable approved drug library with 3D SDF files and docking benchmark datasets, with Load from dataset in molecule inputs
  • Added native SDF output to SMINA, mapped poses to AutoDock-GPU, downloadable design metrics to RFdiffusion3, and native modes, query lists and outputs to ScanNet
  • Added format conversion choices and keyed joins to the workflow editor
  • Added runnable examples to the docs, a Copy page menu, and a Markdown version of every public page at its .md URL
  • Added a Compact/Full toggle for wide result tables and next-step suggestions after downloading results

Improved

  • Improved speed when moving between workspace pages and tools, submitting jobs, searching the tool directory and following job progress
  • Improved scrolling in large molecular dynamics, GC content and open reading frame result tables
  • Improved input checks before submission in IQ-TREE, PRODIGY, ColabDock, AutoDock Vina, AIMNet2, RosettaFold3, RAxML-NG, Clustal Omega and DiffDock
  • Improved error messages in PDB2PQR, MolProbity, PDBFixer, Chai-1, ProteinMPNN and SDF to PDBQT
  • Improved the tool input panel with the job name and project in its header, a wider layout, compact setting tabs and settings shown directly for tools with few options
  • Improved pasted sequence handling, so wrapped sequences are read consistently and FASTA names are kept
  • Improved Caliby settings, model selection and BinaryCIF input, and raised the SMINA maximum runtime to two hours

Fixed

  • Fixed OpenFold 3 predictions with uploaded ligand structures, Genie 3 motif scaffolding on multichain structures and ScanNet submissions with PDB and UniProt identifiers
  • Fixed IQ-TREE rejecting CLUSTAL, PHYLIP and NEXUS alignments, FastTree ignoring zero support resamples, and PROPKA and PDB2PQR rejecting PDB files with charge columns
  • Fixed MOL and SDF files with blank molecule titles failing to convert, and hydrogens being dropped after pH correction in SDF to PDBQT
  • Fixed large electrostatic maps not loading in PDB2PQR results, viewer errors when measuring while removing structures, and file previews showing a previously opened file
  • Fixed chart export dialogs hidden in fullscreen, table statistics overlapping rows, and missing colors in workflow controls and charts
  • Fixed slow or failing workspace loads during busy periods, retrying a tool page after a loading error, and missing results for cancelled jobs fetched through MCP

SMINA adds flexible side-chain docking, and many hosted tools now expose more of their native settings and outputs

Many sequence calculators now handle FASTA descriptions, CSV and PDB inputs more reliably, and several of them return extra results.

Added

Improved

Fixed

Eight new Boltz models and screening tools arrive, alongside DeepRank-Ab, DiffDock-PP and OpenDDE

Every hosted tool now runs on the Free plan, the workflow editor gains undo and pre-run checks, settings move into a searchable dialog, and workspace pages, viewers and tool pages load noticeably faster.

Added

Improved

  • Improved the Free plan so every hosted tool can be run on it, including OpenMM, OpenFE and binder design tools
  • Improved ProteinMPNN and LigandMPNN results with an interleaved alignment view and a Sequences tab for FASTA export
  • Improved the docking pose panel with a ligand filter and per-pose downloads, and made the MSA viewer smoother to scroll
  • Improved input checks before submission for PDBFixer, DSSP, GeoDock, ANARCII, ProLIF, DynamicBind, Chou-Fasman and MD trajectory analysis, and molecule inputs now accept 5-character CCD codes
  • Improved loading speed of 3D structure viewers, tool pages, workspace pages, workflows and the AI Assistant

Fixed

  • Fixed missing results from large Foldseek and RNAplfold runs
  • Fixed FlowDock jobs failing with multiple SMILES ligands and workflow steps rejecting proteins from earlier steps
  • Fixed SASA calculator results for selenomethionine and other modified residues, missing MD trajectory analysis charts, and aliphatic index results for FASTA descriptions and CSV and PDB uploads
  • Fixed run logs not opening for some failed jobs, panels blinking when opening saved results, and a crash when adding residue modifications

CatPred now predicts enzyme kinetics, and FreeBindCraft adds a second hosted option for binder design

FASTA conversion gets two new tools, and more docking and structure tools catch unusable inputs before a run starts.

Added

Improved

  • Improved TXT to FASTA and GenBank to FASTA to recognize more sequence layouts in pasted text and documents
  • Improved DNA Shuffle with uniform shuffling and support for messy input, RNA and document uploads, and made seeded runs in the DNA mutation generator consistent
  • Improved fpocket so residue-defined pockets can be selected directly and structures with no detected pockets still return results
  • Improved AlphaFold2 support for very long inputs and kept the original error details when a prediction fails
  • Improved input checks before submission for Boltz-2 templates, PLIP, AutoDock Vina search boxes and multi-model receptors, AutoDock-GPU receptors, ViennaRNA, Aggrescan3D and PPAP
  • Improved speed of 3D pose viewing, batch validation and search, CSV previews and workflow runs
  • Improved input size messages on the free plan and made alignment format checks consistent across tools

Fixed

  • Fixed file previews losing their original formatting
  • Fixed workflow ligand inputs that rejected some supported file formats

Batches now accept CSV files that describe each job in full, and design and docking tools share a residue picker that works directly on the structure

Workflows also got a clearer vertical layout.

Added

  • Added CSV batches for named complexes, explicit file pairings and settings that vary between jobs
  • Added a shared residue picker to AutoDock Vina, ProteinMPNN, LigandMPNN, SolubleMPNN and Genie 3 for selecting residues on the structure preview
  • Added a way to report an input validation message you think is wrong directly from the input card
  • Added the option to hide the app sidebar, with a preview on hover

Improved

  • Improved workflows with a vertical layout, more reusable outputs, and clearer explanations when a credit estimate cannot be calculated
  • Improved AutoDock Vina to repair incomplete receptors automatically, including repaired structures passed in from workflows
  • Improved FlowDock with the shared structure viewer and faster input preparation
  • Improved input checks before submission for PepMimic, ProteinMPNN and PROPKA 3, and gave PROPKA more time for large structures
  • Improved Salmon to accept short reads

Fixed

  • Fixed 3D structure display for single molecules in SMILES to PDB and related SMILES converters
  • Fixed some valid structure and ligand files being rejected at upload

Batches can now pair inputs in order instead of running every combination, and ORF Finder and BioPhi results are easier to explore

Large tables, alignments and file libraries also load faster.

Added

  • Added a pairwise option to batches that matches inputs from separate lists in order instead of running all combinations
  • Added explorer, table and summary views to ORF Finder, with linked ORF selection and zoom
  • Added native geometry settings to PeptideBuilder and refinement controls to GeoDock

Improved

  • Improved BioPhi results with grouping by chain and sequence variant, and input checks before submission
  • Improved Admetica with grouped property selection
  • Improved speed of large result spreadsheets, the alignment viewer, large file libraries and 3D structure loading
  • Improved input checks before submission for DSSP, USAlign, Boltz-2 bonds and AutoDock Vina mmCIF ligands and flexible residues
  • Improved DSSP to explain empty result tables and PDB2PQR to show its run summary separately from pKa values

Fixed

  • Fixed batch downloads missing some result files
  • Fixed AutoDock Vina dropping some uploaded ligands
  • Fixed AlphaFlow showing selectors for inactive models

Ramachandran plot now makes individual residues easier to inspect, reports analysis coverage, and keeps filtered statistics and exports aligned with the plot

Added

  • Added a ready-to-run 1HNR example, direct PDB ID entry, keyboard and touch residue inspection, and an outliers-only filter to Ramachandran plot
  • Added per-residue skip reasons and separate current-selection and whole-structure CSV exports

Improved

  • Improved plot readability in narrow output panels with equal angle scales, separate residue-class reference views, degree labels, and numerical summaries

Fixed

  • Fixed angle calculation for incomplete neighboring residues, undefined geometry, and twisted-proline reference selection