Branch: refs/heads/ehb54-issue-1098 Home: https://github.com/ehb54/ultrascan3 Commit: e2516443a7400573e18cbf67559d6024a235cab8 https://github.com/ehb54/ultrascan3/commit/e2516443a7400573e18cbf67559d6024a... Author: ehb54 <brookes@uthscsa.edu> Date: 2026-10-01 (Thu, 01 Oct 2026) Changed paths: M us_somo/develop/src/us_hydrodyn_hydrate_align.cpp M us_somo/develop/src/us_hydrodyn_pdb_tool_merge.cpp M us_somo/develop/src/us_saxs_util_hydrate_align.cpp M us_somo/somo/doc/manual/somo/somo_pdb_editor_merge.html Log Message: ----------- somo: fix the atom_align reflection case and seven Cut/Splice defects atom_align, in both its US_Hydrodyn and US_Saxs_Util copies, corrected a reflection, det( V U^T ) < 0, with rot = U S V^T. The Kabsch correction is R = V S U^T, so the code applied the inverse of the correct rotation. Three points are always coplanar, so with three alignment points this branch is taken about half of the time: - Cut/Splice: Compute Guess and Only Closest Chains propose fits on three residues. In tests from random donor orientations, 5 of 10 splices misplaced the merged residues by 2 to 32 A RMSD, and 5 of 6 Only Closest Chains runs misplaced an added chain by up to 130 A. After the fix, 0 of 10 and 0 of 6. - Hydration: waters are positioned on up to four atoms (three for main chain waters), and coplanar sets (main chain, carboxylate, amide, guanidinium, aromatic rings) take the same branch. Hydrating RNase A (8RAT) now places 82 waters instead of 73. The 19 waters that move are replaced by 28, of which 23 (82%) lie within 2.5-3.3 A of a polar atom, against 12 of 19 (63%) before. Cut/Splice (us_hydrodyn_pdb_tool_merge.cpp): - Extra Chains blanked the chain name of the rows it created, because table_value() ran on each setItem(). Disconnect while filling, as Only Closest Chains already does. - Loading a .csc row ending in an empty value, which is how Save writes a row without a cut, left a table cell undefined, and SOMO crashed at the next Validate, Trial, Start or Save. Pad rows to the full width, fill the table without running table_value() (which also blanked the chain of + rows), and reject an empty file. - Start with the steric clash filter crashed on a Chains To file without MODEL records when a clash was found (model name lookup in an empty vector). - Start never finished when three models shared a MODEL number: the -N suffix of the duplicate name never advanced. - Cancelling the Target dialog set the target to ".pdb". - From -> Merge, From -> Fit, To -> Fit and To -> Cut adjusted the row at the chain's position in the Chains To structure instead of the chain's row in the table. - Start dropped the last line of a Chains To file without a final END. Update the manual page accordingly: the defect notes are removed, and widening the fit range remains as advice. Addresses items 1-8 of ehb54/ultrascan-tickets#1097. Item 9 (no automatic adjustments for loaded rows) and the clash filter tolerance remain open. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Commit: ab7fb5909b4370635d7927ddb1c970bc962906c3 https://github.com/ehb54/ultrascan3/commit/ab7fb5909b4370635d7927ddb1c970bc9... Author: emre brookes <ehb54@users.noreply.github.com> Date: 2026-10-01 (Thu, 01 Oct 2026) Changed paths: M us_somo/develop/src/us_hydrodyn_hydrate_align.cpp M us_somo/develop/src/us_hydrodyn_pdb_tool_merge.cpp M us_somo/develop/src/us_saxs_util_hydrate_align.cpp M us_somo/somo/doc/manual/somo/somo_pdb_editor_merge.html Log Message: ----------- Merge pull request #576 from ehb54/ehb54-issue-1097 somo: fix the atom_align reflection case and seven Cut/Splice defects Compare: https://github.com/ehb54/ultrascan3/compare/cbf1713c78a8...ab7fb5909b43 To unsubscribe from these emails, change your notification settings at https://github.com/ehb54/ultrascan3/settings/notifications