[us-commits] [ehb54/ultrascan3] 95fec1: somo/doc: document GRPY shell reduction in the Hyd...

[emre brookes]

  Branch: refs/heads/ehb54-issue-984
  Home:   https://github.com/ehb54/ultrascan3
  Commit: 95fec1cc63cc817cc419d9e1eeb4027a9793e0ad
      https://github.com/ehb54/ultrascan3/commit/95fec1cc63cc817cc419d9e1eeb4027a9793e0ad
  Author: ehb54 <brookes at uthscsa.edu>
  Date:   2026-08-06 (Thu, 06 Aug 2026)

  Changed paths:
    M us_somo/somo/doc/manual/somo/somo_hydro.html

  Log Message:
  -----------
  somo/doc: document GRPY shell reduction in the Hydrodynamics Options manual

Adds a section to somo_hydro.html covering the new "GRPY Shell Reduction"
box, placed after the GRPY buried-bead exclusion it generalizes, and bumps
the page's Last updated stamp.

Covers what the procedure does (repeated exact GRPY on progressively larger
subsets of the most solvent-exposed beads, stopping when successive results
agree, with the remaining error estimated by Richardson extrapolation), the
three controls and their defaults, and why the check is nearly free.

Gives particular attention to the "Require intrinsic viscosity" checkbox,
since the hydrodynamic quantities do not tolerate reduction equally: at equal
reduction the rotational diffusion error is ~1.8x the translational, and the
intrinsic viscosity ~3.3x and systematically one-sided. It is the viscosity
that determines when the procedure can stop, so requiring it costs much of
the speed gain; when not required, the viscosity and the Einstein radius
derived from it are withheld from the results rather than reported with a
caveat, while being retained and annotated in the results file.

Also answers, quantitatively, the question the preceding buried-bead
paragraph explicitly left open ("would need an in depth investigation",
supported only by preliminary 2021 tests on lysozyme): retaining >50% of
beads changes the translational diffusion coefficient by ~0.02%, 30-50% by
~0.12%, 20-30% by ~0.5%, degrading rapidly below ~10%. Buried-bead exclusion
typically retains ~40%, so that paragraph's regime is now measured. Notes
the reported error bounded the true deviation in all 344 validation tests.

Ends with the honest practical caveat: the gain grows with model size,
is negligible for small models, and on top of the default buried-bead
exclusion is typically 2-3x rather than the larger factors available from a
completely unreduced model -- the lasting benefit being the error estimate,
not the speed.

All markup uses named entities, so the page remains pure ASCII under its
declared ISO-8859-1 charset.

Fixes ehb54/ultrascan-tickets#984

Co-Authored-By: Claude Opus 5 <noreply at anthropic.com>



To unsubscribe from these emails, change your notification settings at https://github.com/ehb54/ultrascan3/settings/notifications