@platforma-open/milaboratories.3d-structure-prediction 1.2.3 → 1.3.2

package/block-pack/CHANGELOG.md

@@ -0,0 +1,3 @@
+## 1.0.0
+
+Initial release.

package/block-pack/description.md

@@ -0,0 +1,17 @@
+# Overview
+
+Predicts 3D structures of antibody variable domains directly from sequence, producing per-clonotype PDB files with residue-level confidence scores.
+
+The block uses [ImmuneBuilder](https://github.com/oxpig/ImmuneBuilder) — a deep-learning model trained on the SAbDab database that runs an ensemble of four predictors per antibody. **ABodyBuilder2** handles conventional paired antibodies (VH + VL) and is selected automatically when a heavy and a light chain column are both configured; **NanoBodyBuilder2** handles VHH nanobodies and is selected when only a heavy chain is configured. ImmuneBuilder is fast (orders of magnitude faster than AlphaFold-Multimer for antibodies) and produces models comparable in CDR-H3 accuracy to experimental crystal structures.
+
+Confidence is reported as the **per-residue ensemble disagreement in Ångströms** — the standard deviation of Cα positions across the four ensemble members. Lower values mean higher confidence; framework regions typically resolve to under 1 Å, while CDR-H3 (the most challenging loop) ranges from ~1 Å for short loops to several Å for long, flexible ones. The block exports both whole-structure and per-CDR aggregates, so users can filter on the loop quality that matters for their downstream task. A user-configurable confidence threshold (default 2.5 Å on CDR-H3) flags benchmark-quality predictions and produces a `confident` subset column for direct use in lead selection or clustering.
+
+Predicted structures are IMGT-numbered with CDR boundary annotations baked into each PDB. The block also tracks per-row outcomes — sanitization rejections, signal-peptide warnings, long-CDR-H3 flags, missing VHH hallmarks, and prediction failures — surfaced as plain-text "Failure reason" and "Warnings" columns so quality issues are visible in the table rather than silently corrupting downstream analysis.
+
+ImmuneBuilder is developed by the [Oxford Protein Informatics Group](https://www.stats.ox.ac.uk/research/oxford-protein-informatics-group). For more information, please see: [https://github.com/oxpig/ImmuneBuilder](https://github.com/oxpig/ImmuneBuilder) and cite the following publication if used in your research:
+
+> Abanades B, Wong WK, Boyles F, Georges G, Bujotzek A, Deane CM. *ImmuneBuilder: Deep-Learning models for predicting the structures of immune proteins.* Communications Biology, 6, 575 (2023). [https://doi.org/10.1038/s42003-023-04927-7](https://doi.org/10.1038/s42003-023-04927-7)
+
+Residue numbering uses [ANARCI](https://github.com/oxpig/ANARCI):
+
+> Dunbar J and Deane CM. *ANARCI: antigen receptor numbering and receptor classification.* Bioinformatics, 32(2):298-300 (2016). [https://doi.org/10.1093/bioinformatics/btv552](https://doi.org/10.1093/bioinformatics/btv552)

package/block-pack/manifest.json

@@ -0,0 +1 @@
+{"schema":"v2","description":{"id":{"organization":"milaboratories","name":"3d-structure-prediction","version":"1.3.2"},"components":{"workflow":{"type":"workflow-v1","main":{"type":"relative","path":"main.plj.gz"}},"model":{"type":"relative","path":"model.json"},"ui":{"type":"relative","path":"ui.tgz"}},"meta":{"title":"3D Structure Prediction","description":"Predicts antibody and VHH 3D structures with ImmuneBuilder. Emits IMGT-numbered per-clonotype PDB files with per-residue confidence.","longDescription":{"type":"relative","path":"description.md"},"changelog":{"type":"relative","path":"CHANGELOG.md"},"logo":{"type":"relative","path":"block-logo.png"},"url":"https://github.com/platforma-open/milaboratories.3d-structure-prediction","support":"mailto:support@milaboratories.com","tags":["airr","downstream","antibody","structure"],"organization":{"name":"MiLaboratories Inc","url":"https://milaboratories.com/","logo":{"type":"relative","path":"organization-logo.png"}},"supportedPlatforms":["linux-x64","macosx-aarch64"]},"featureFlags":{"supportsLazyState":true,"supportsPframeQueryRanking":true,"requiresUIAPIVersion":3,"requiresModelAPIVersion":2,"requiresCreatePTable":2,"requiresPFramesVersion":1001031,"requiresPFrameSpec":true,"requiresPFrame":true,"requiresDialog":true,"requiresColumnsCollection":true},"kind":"@platforma-open/milaboratories.3d-structure-prediction.kind@1.0.0"},"timestamp":1788168863978,"files":[{"name":"main.plj.gz","size":486197,"sha256":"4A1F5C8DA91A0B9188A5A104EF1C64D175312D9DD737EC4057D7A20831283B69"},{"name":"model.json","size":559230,"sha256":"3B1C23CFB0C00EF37AC318E28BD0B25A59413C09383004F6D2C64823D3BE4DE8"},{"name":"ui.tgz","size":9071573,"sha256":"3A32CC99A636500E8A2B35FCC2C884B83F4429FE68582C59E25A8584A3416E31"},{"name":"organization-logo.png","size":24439,"sha256":"FA71390C77C91E4B7FAAE5640D00F92F1E3F2869296F68B6040DD7CC549A50B5"},{"name":"description.md","size":2780,"sha256":"6E3D18A2D6C920AC4241F67AA2A580895D8F067DA354D9F02E0FAFA5AED6D84C"},{"name":"CHANGELOG.md","size":27,"sha256":"3E1D1DA422E5D609FCDE7F996A81F7635011F3CA297923598DB36D924C4A0DAE"},{"name":"block-logo.png","size":21527,"sha256":"6BB33BAF0CD039549661B51AE490373BE60D1811EC71F5023400928293BC2427"}]}