Pith. sign in

REVIEW 4 major objections 6 minor 14 references

Towards More Accurate Full-Atom Antibody Co-Design

T0 review · 4 major / 6 minor · reviewed 2026-08-08 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read Igformer, an end-to-end full-atom antibody co-design model, claims new best results on CDR-H3 sequence recovery and antibody-antigen docking by refining the binding-interface graph with personalized propagation plus global attention.

desk verdict Plausible incremental model for antibody co-design, but the headline performance claim rests on an unverified train/test split and the equivariance theorem appears to break on the actual pipeline. read the letter →

arxiv 2502.19391 v1 pith:FIUIV7DY submitted 2025-02-11 q-bio.BM cs.LG

Jiayang Wu , Xingyi Zhang , Xiangyu Dong , Kun Xie , Ziqi Liu , Wensheng Gan , Sibo Wang , Le Song This is my paper · ORCID
classification q-bio.BMcs.LG
keywords antibodyco-designcomplementarity-determiningregionCDR-H3equivariantgraphneuralnetworkfull-atomstructurepredictionantigen-bindinginterfaceaminoacidrecoveryDockQ
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper tries to establish that antibody co-design—jointly predicting the amino acid sequence and the 3D structure of the antibody loops that bind an antigen—can be made substantially more accurate by modeling the antibody-antigen binding interface as a graph refined at two scales. It introduces Igformer, an end-to-end full-atom model whose interface update first propagates local chemical and geometric information with personalized propagation, then captures global dependencies with transformer-style attention, and uses the refined interface to drive sequence and structure prediction. The reported payoff is concrete: on the RAbD CDR-H3 task Igformer reaches 43.50% amino acid recovery, 7.15 Å RMSD, and DockQ 0.450, compared with dyMEAN's 42.64%, 8.42 Å, and 0.408, and on the IgFold complex-structure benchmark it reaches DockQ 0.522 versus 0.452. If these numbers hold, multi-scale interface refinement of this kind would give computational antibody design a better starting point for therapeutic candidates, improving both the sequence and the docking quality without a separate docking stage.

What carries the argument

The load-bearing mechanism is the two-scale inter-graph refinement between the antibody paratope and the antigen epitope. An Equivariant Message Passing (EMP) module, which updates residue embeddings and atomic coordinates while preserving $E(3)$-equivariance, supplies initial residue representations. The inter-graph is then refined by Approximate Personalized Propagation (APP), which balances local neighborhood context against initial residue features, and by a Simplified Graph Transformer (SGFormer), which applies global attention across the whole binding interface. The refined inter-graph is passed through a Triangle Multiplicative module and Axial Attention module, and finally through separate EMP modules for intra- and inter-graph processing that generate the predicted coordinates and amino acid sequences. The paper proves that coordinates are $E(3)$-equivariant and embeddings are $E(3)$-invariant, and its ablations identify the inter-graph refinement and dual-scale message passing as the main sources of the reported DockQ improvement.

What would settle it

Compare every one of the 3,246 training antibodies with each of the 60 RAbD and 51 IgFold test complexes by CDR-H3 sequence identity at the 40% threshold used for clustering; any test complex that lands in a training cluster would refute the leakage-free evaluation claim.

Watch

Extended reading notes

Core claim

On its own terms, the paper's central discovery is that a carefully refined antibody-antigen inter-graph converts interface modeling into a two-scale problem: local chemical and geometric constraints are propagated by Approximate Personalized Propagation, long-range dependencies are captured by a Simplified Graph Transformer, and the refined edge structure then drives separate intra- and inter-graph equivariant message passing for full-atom structure and sequence prediction. The paper claims this yields state-of-the-art epitope-binding CDR-H3 co-design on the RAbD benchmark, with amino acid recovery 43.50% and DockQ 0.450 versus dyMEAN's 42.64% and 0.408, and stronger complex structure prediction on the IgFold benchmark, with DockQ 0.522 versus 0.452. Ablation studies in the paper attribute the gain to each component, with the personalized propagation and the dual EMP design showing the largest drops when removed.

Load-bearing premise

The load-bearing premise is that the 60 RAbD and 51 IgFold test complexes are genuinely absent from the 3,246-antibody SAbDab training set; the paper asserts this prevents leakage but reports no overlap check, so a hidden match at the 40% CDR-H3 clustering threshold would turn the reported gains into a memorization artefact.

Editorial extensions

If this is right

  • CDR-H3 co-design improves on both sequence and docking fronts: 43.50% amino acid recovery, 7.15 Å RMSD, and DockQ 0.450 on RAbD, each ahead of dyMEAN.
  • Designing all six CDRs also improves (63.55% versus 60.05% AAR, and DockQ 0.4817 versus 0.3973), while full-antibody design improves less but consistently (73.69% AAR and DockQ 0.4600).
  • On complex structure prediction from complete antibody sequences, Igformer reaches DockQ 0.522 versus dyMEAN's 0.452 on the IgFold benchmark, and the paper reports it also outperforms a ground-truth-guided HERN upper bound.
  • Ablations show that every proposed component contributes, with removing APP causing the largest DockQ drop on Task 1 (0.416 versus 0.450) and removing the dual EMP design being close behind.
  • Because the coordinates are $E(3)$-equivariant and the embeddings are $E(3)$-invariant, predictions are consistent under rigid rotations and translations, a physical requirement for generated antibody structures.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Editorial inference: if the leakage-free split holds, the model should generalize to epitopes deposited after the November 2022 SAbDab snapshot; testing on a time-split of newer complexes would separate learned binding physics from training-set familiarity.
  • Editorial inference: the design principle of local propagation plus global attention over an inter-graph is not antibody-specific and could plausibly transfer to nanobodies or T-cell receptor interfaces, though the paper does not test this.
  • Editorial inference: even at 43.5% CDR-H3 recovery, most residues in the most variable loop are not recovered, so practical therapeutic use would likely pair Igformer with affinity maturation or experimental screening, a step the paper leaves implicit.
  • Editorial inference: because the reported gains are measured on docking metrics against the full antigen, the interface loss's 6.6 Å contact threshold may understate long-range or conformational binding contributions; ablating that threshold would clarify whether the DockQ gain comes from local contact geometry or global paratope placement.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 6 minor

Summary. The manuscript proposes Igformer, an end-to-end equivariant graph network for full-atom antibody CDR co-design. It represents antibody-antigen complexes as intra- and inter-graphs, uses Approximate Personalized Propagation and a simplified graph transformer to refine the paratope-epitope interface, and applies triangle multiplicative and axial attention modules before dual equivariant message passing for coordinate and sequence prediction. Evaluation on the RAbD and IgFold benchmarks reports improved AAR, RMSD, and DockQ over dyMEAN and earlier baselines across CDR-H3 design, multiple CDR design, full antibody design, and complex structure prediction.

Significance. If the empirical results survive a leakage-free evaluation, Igformer would represent a meaningful advance in antibody co-design, improving both sequence recovery and docking quality over the previous end-to-end state of the art. The proposed interface-refinement architecture is plausible, and the ablation studies consistently show contributions from APP, SGFormer, triangle modules, and dual EMP. However, the headline comparisons currently rest on an asserted but unverified separation of training and test sets, on baselines whose training conditions are not shown to be matched, on missing uncertainty estimates, and on an equivariance theorem that is proved only for the EMP submodule. These gaps must be closed before the central claims can be accepted.

major comments (4)
  1. [Section 2.1 and Appendix E.1] Section 2.1 states that 'This test set selection prevents data leakage during the evaluation phase,' but no overlap analysis is reported between the 3,246 training antibodies and the 60 RAbD or 51 IgFold test complexes. Appendix E.1 explicitly says that the curated SAbDab dataset 'serves as the foundation for both the RAbD and the IgFold benchmark dataset,' so the test complexes are drawn from the same source snapshot used for training. Because the training set is clustered at 40% CDR-H3 identity, the decisive check is an identity and structural overlap report after that clustering; without it, the reported gains in AAR, RMSD, and DockQ in Tables 1-4 could reflect memorization rather than generalization. Please report overlap statistics and remove any training antibodies within the clustering threshold of a test antibody, or justify why this is unnecessary.
  2. [Appendix E.3 and Tables 1-4] Tables 1-4 report single-point performance values without error bars, and Appendix E.3 says that baselines are run with 'hyperparameters and training procedures from their official releases' rather than retrained on the same SAbDab train/validation split used for Igformer. The caption of Table 1 further says that results with * are 'collected from the dyMEAN paper,' which is not the same as reproducing those baselines under identical conditions. To support the headline comparisons, including AAR 43.50 versus 42.64, RMSD 7.15 versus 8.42, and DockQ 0.450 versus 0.408 in Table 1, the authors should either retrain all baselines on the identical split with multiple seeds or provide explicit evidence that the published numbers were obtained on the same snapshot and split, and they should report means and standard deviations over seeds or checkpoints.
  3. [Theorem C.3 and Appendix D.5] Theorem C.3 claims E(3)-equivariance of the full Igformer pipeline, but the proof in Appendix D.5.2-D.5.4 only shows that the EMP coordinate update is equivariant under rotations and translations of a fixed set of input coordinates. It does not cover the template-based coordinate initialization and Kabsch alignment of Appendix D.3.2, the chain-wise centering and dimension-specific normalization of Appendix D.3.3, or the iterative replacement of epitope and paratope coordinates in Algorithm 1. The normalization step Xi = (X'_i - mu_d)/sigma_d is not an E(3) transformation in general, so Theorem C.3 is not established as stated. Either extend the proof to the full pipeline or state the equivariance claim only for the EMP module.
  4. [Appendix F.1 and F.2] Appendix F.1 (Table 11) selects the paratope definition per task, using H3 for Tasks 1, 3, and 4 and H3+L3 for Task 2, and Appendix F.2 (Figure 5) chooses w = 0.2 as the value with peak DockQ across all tasks. If these choices were made after observing test benchmark results, the reported performance is optimistic and the comparisons to dyMEAN are not on equal footing. Please state whether these hyperparameters were fixed before evaluating on RAbD and IgFold, and describe the validation-set selection procedure.
minor comments (6)
  1. [Abstract and Section 2.5] The abstract reports an 11.84% reduction in RMSD, while Section 2.5 reports a 12.44% relative improvement in RMSD; these numbers should be reconciled.
  2. [Section 2.2 and Abstract] The text says Igformer achieves a 2.2% relative improvement in AAR over dyMEAN, while the abstract and Table 1 imply a relative improvement of 2.02% (43.50/42.64 - 1); please make the description consistent.
  3. [Table 1 caption] The caption states that results with * are collected from the dyMEAN paper, but the table itself contains no star symbols; please insert the markers or remove the note.
  4. [Appendix E.1] The sentence referring to the RAbD benchmark cites Dunbar et al. (2013), but the RAbD benchmark is defined by Adolf-Bryfogle et al. (2018); the citation should be corrected.
  5. [Appendix D.3.3] The final sentence says that 'we update the global coordinates again based on Equations 12 and 13,' which is too vague to determine whether normalization is reapplied; please specify the exact coordinate transformation used at this stage.
  6. [Table 8 caption] The caption contains a typo, 'inter-grpah refinement module,' which should read 'inter-graph refinement module.'

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the headline results are external-benchmark evaluations and the equivariance proofs are structural, not predictive, so the derivation does not reduce to its inputs.

full rationale

The paper's central claims (Tasks 1-4) are evaluated on external benchmarks, RAbD and IgFold, using AAR, RMSD, TM-score, lDDT, and DockQ. The sequence and structure predictions are produced by the model from learned embeddings and coordinates (Section C.6), trained with cross-entropy and Huber losses (Equations 16-19) against ground-truth PDB structures. The reported test-set numbers are therefore independent evaluations, not algebraic consequences of the training inputs or of the loss function itself. The E(3)-equivariance claims (Theorems C.1, C.3, D.4, D.9) are proven by induction from the explicit EMP update rules, which are architectural properties rather than empirical predictions; they do not assert that any particular antibody structure is correct. No load-bearing self-citation appears: the references to SAbDab, dyMEAN, RAbD, IgFold, and the personalized-propagation literature are external prior work by different author groups, and the template-residue threshold is adopted from dyMEAN as an external convention rather than used to forbid alternatives. The hyperparameter analysis (w = 0.2 in Figure 5, and the H3 vs. H3+L3 paratope comparison in Table 11) is ordinary model selection; importantly, the paper reports both paratope variants and chooses per task, so a fitted choice is not being relabeled as a prediction. The concern raised by the reviewer that RAbD and IgFold complexes may overlap the SAbDab training snapshot is a data-leakage/correctness risk, not a circularity: the paper asserts that its split 'prevents data leakage' without showing a sequence-identity overlap analysis, but this is an unsupported empirical claim about benchmark hygiene, not a step in which an output is defined in terms of an input or in which a fitted parameter is renamed as a prediction. Accordingly, no specific circular step can be quoted and demonstrated, and the appropriate finding is no significant circularity (score 0).

Assumptions & free parameters 7 free parameters · 5 assumptions · 0 invented entities

The central claim is an empirical performance claim, so the ledger lists the hand-set hyperparameters that shape the model and the domain assumptions the evaluation depends on. The most load-bearing assumptions are the absence of test leakage and the validity of the in-silico metrics as proxies for antibody function.

free parameters (7)
  • Similarity blending weight w = 0.2
    Blends residue-level and atom-level similarities in the EMP module (Section C.2); Figure 5 shows DockQ peaks at w=0.2 across all tasks, so the value was selected by scanning.
  • APP retention weight alpha = 0.1
    Controls how much initial node features are retained versus neighbor propagation in Eq. (1) (Section C.3); set in Table 8 without sensitivity analysis.
  • SGFormer residual weight alpha = 0.9
    Residual weight in the simplified graph transformer (Table 8); emphasizes transformed features and is fixed by the authors.
  • kNN neighbor count k = 9
    Number of nearest neighbors for intra- and inter-graph edge construction (Appendix E.2). Affects the receptive field of message passing.
  • Contact distance threshold = 6.6 Å
    Defines interacting residues for interface loss, contact AAR, and epitope-paratope edges (Appendix D.7, E.4); a standard but hand-chosen cutoff.
  • Epitope size = 48 residues
    Top-k antigen residues with smallest distance to CDR-H3 form the epitope (Appendix D.4); fixed for computational efficiency.
  • Conserved residue template threshold = 95% conservation; 16 heavy + 18 light positions
    Criterion for template backbone coordinates (Appendix D.3.1), described as 'optimized through empirical analysis' and inherited from dyMEAN.
assumptions (5)
  • domain assumption The RAbD and IgFold test complexes are absent from the SAbDab training snapshot at the clustering threshold.
    Section 2.1 claims the test set selection prevents leakage, but no overlap analysis is provided. If false, the central performance claim collapses.
  • domain assumption The coordinate preprocessing in D.3.3 is compatible with the E(3)-equivariance theorem.
    Theorem C.3 assumes inputs transform as X -> QX + b, but D.3.3 applies chain-specific centering and per-dimension normalization, which is not rotation-equivariant for raw structures. The theorem holds only for the normalized coordinates, not the end-to-end mapping.
  • domain assumption AAR, RMSD, TM-score, lDDT, and DockQ are meaningful proxies for antibody binding function.
    All significance claims rely on these in-silico metrics; no wet-lab validation is included.
  • domain assumption Template coordinates built from the mean of 16/18 conserved framework residues generalize to unseen antibodies.
    Appendix D.3.1 assumes the mean backbone coordinates of conserved residues are a suitable initialization for any new antibody.
  • standard math Standard properties of MLPs, attention, Kabsch alignment, and linear algebra are used without proof.
    Relied upon in the proofs of equivariance and in coordinate initialization.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Towards More Accurate Full-Atom Antibody Co-Design." pith.science (2026). https://pith.science/paper/FIUIV7DY

@misc{pith2026250219391,
  author       = {Pith},
  title        = {Pith review of: Towards More Accurate Full-Atom Antibody Co-Design},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/FIUIV7DY}},
  note         = {Machine review of arXiv:2502.19391}
}
read the original abstract

Antibody co-design represents a critical frontier in drug development, where accurate prediction of both 1D sequence and 3D structure of complementarity-determining regions (CDRs) is essential for targeting specific epitopes. Despite recent advances in equivariant graph neural networks for antibody design, current approaches often fall short in capturing the intricate interactions that govern antibody-antigen recognition and binding specificity. In this work, we present Igformer, a novel end-to-end framework that addresses these limitations through innovative modeling of antibody-antigen binding interfaces. Our approach refines the inter-graph representation by integrating personalized propagation with global attention mechanisms, enabling comprehensive capture of the intricate interplay between local chemical interactions and global conformational dependencies that characterize effective antibody-antigen binding. Through extensive validation on epitope-binding CDR design and structure prediction tasks, Igformer demonstrates significant improvements over existing methods, suggesting that explicit modeling of multi-scale residue interactions can substantially advance computational antibody design for therapeutic applications.

Figures

Figures reproduced from arXiv: 2502.19391 by the authors.

Figure 1
Figure 1. End-to-end antibody co-design [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Framework of Igformer. (2021); Jin et al. (2022b); Watson et al. (2022b); Luo et al. (2022); Kong et al. (2023a); Martinkus et al. (2023); Wang et al. (2024a), have emerged as powerful tools for simultaneously co-designing CDR 1D sequences and 3D structures. These models represent a significant advancement over traditional approaches by leveraging deep learning to model complex sequential and structural relationship… view at source ↗
Figure 3
Figure 3. Antibody structures by Igformer & dyMEAN. [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: Illustration of antibody-antigen complex. [PITH_FULL_IMAGE:figures/full_fig_p012_4.png]
Figure 1
Figure 1. Figure 1: We formulate this problem using a dual-graph representation of the binding region. The [PITH_FULL_IMAGE:figures/full_fig_p012_1.png]
Figure 5
Figure 5. Figure 5: DockQ scores for varying w values across four tasks, with the peak observed at w = 0.2, emphasizing the importance of geometric distance in similairty matrix computation. achieves its peak at w = 0.2 across all tasks. This optimal value indicates that atom-level simila…
Figure 6
Figure 6. Figure 6: Additional antibody structures generated by Igformer. [PITH_FULL_IMAGE:figures/full_fig_p030_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

14 extracted references · 11 canonical work pages

  1. [1]

    The first component is an antibody graph Gab = {Vab, Eab}, while the second is an antigen epitope graph Gae = {Vae, Eae}

    We formulate this problem using a dual-graph representation of the binding region. The first component is an antibody graph Gab = {Vab, Eab}, while the second is an antigen epitope graph Gae = {Vae, Eae}. The vertices Vab and Vae represent amino acid residues, where each residue vi is characterized by its amino acid type si and a coordinate matrix Xi ∈ R3...

    work page 1976

  2. [3]

    (2019) and train the model on a single GeForce RTX 3090 Ti GPU using the Adam optimizer Kingma (2015)

    NodeMLP Linear(64 → 128), SiLU(), Dropout(0.1) Linear(128 → 128), SiLU(), Dropout(0.1) E.2 I MPLEMENTATION DETAILS We implement Igformer using PyTorch Paszke et al. (2019) and train the model on a single GeForce RTX 3090 Ti GPU using the Adam optimizer Kingma (2015). The model training process utilizes a batch size of 16, running for 150 epochs in task 1 ...

    work page 2019

  3. [7]

    Iggm: A generative model for functional antibody and nanobody design

    Rubo Wang, Fandi Wu, Xingyu Gao, Jiaxiang Wu, Peilin Zhao, and Jianhua Yao. Iggm: A generative model for functional antibody and nanobody design. bioRxiv, 2024a. Shixia Wang, Kun-Wei Chan, Danlan Wei, Xiuwen Ma, Shuying Liu, Guangnan Hu, Saeyoung Park, Ruimin Pan, Ying Gu, Alexandra F Nazzari, et al. Human cd4-binding site antibody elicited by polyvalent ...

    work page 2022

  4. [13]

    18 D.3.4 P ARATOPE COORDINATES GENERATION Our paratope coordinate generation framework implements a structured approach to initialize and refine binding interface positions. The process begins by anchoring initial paratope coordinates to the epitope center X0 = Xc where X0 represents the initial paratope coordinates and Xc represents the coordinates of th...

    work page 1989

  5. [14]

    (2014) mechanisms after activation functions

    Each layer incorporates dropout Srivastava et al. (2014) mechanisms after activation functions. The similarity features are normalized after being mapped from 196 to 128 dimensions, including simres ij and simatom ij . The iterative application of these components ensures efficient updates of both node features and coordinate transformations while maintai...

    work page 2014

  6. [20]

    We save model parameters from the top 10 validation rounds, compute metrics for each model independently, and report the mean values as final results

    E.3 B ASELINES For the baselines, we adopt the hyperparameters and training procedures from their official releases since all methods utilize SAbDab to construct training sets of similar scale and distribution. We save model parameters from the top 10 validation rounds, compute metrics for each model independently, and report the mean values as final resu...

    work page 2018

  7. [48]

    This selection process ensures that we focus on the most critical residues involved in antibody-antigen binding while maintaining computational efficiency

    smallest distances to form the antigen epitope: Vae = {arg min k ˆdist(vi) | ∀vi ∈ Vag}. This selection process ensures that we focus on the most critical residues involved in antibody-antigen binding while maintaining computational efficiency. The fixed-size antigen epitope selection provides a consistent representation of the binding interface, enabling...

    work page 1976

  8. [1989]

    Dreyer, Daniel Cutting, Constantin Schneider, Henry Kenlay, and Charlotte M

    Fr´ed´eric A. Dreyer, Daniel Cutting, Constantin Schneider, Henry Kenlay, and Charlotte M. Deane. Inverse folding for antibody sequence design using deep learning. CoRR, abs/2310.19513,

    arXiv

Show all 14 references
  1. [2005]

    12 C.2 Equivariant Message Passing

    9 Table of Contents A Related Work 11 B Problem Formulation 11 C Method 12 C.1 Initialization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 C.2 Equivariant Message Passing . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 C.3 Inter-Graph Ref...

    2022

  2. [2016]

    Deep learning methods for designing proteins scaffolding functional sites

    Jue Wang, Sidney Lisanza, David Juergens, Doug Tischer, Ivan Anishchenko, Minkyung Baek, Joseph L Watson, Jung Ho Chun, Lukas F Milles, Justas Dauparas, et al. Deep learning methods for designing proteins scaffolding functional sites. BioRxiv, pp. 2021–11,

    2021

  3. [2018]

    Protein generation with evolutionary diffusion: sequence is all you need

    Sarah Alamdari, Nitya Thakkar, Rianne van den Berg, Neil Tenenholtz, Bob Strome, Alan Moses, Alex Xijie Lu, Nicolo Fusi, Ava Pardis Amini, and Kevin K Yang. Protein generation with evolutionary diffusion: sequence is all you need. BioRxiv, pp. 2023–09,

    2023

  4. [2021]

    Dropout: a simple way to prevent neural networks from overfitting

    Nitish Srivastava, Geoffrey Hinton, Alex Krizhevsky, Ilya Sutskever, and Ruslan Salakhutdinov. Dropout: a simple way to prevent neural networks from overfitting. JMLR, 15(1):1929–1958,

    1929

  5. [2022]

    Progen: Language modeling for protein generation

    Ali Madani, Bryan McCann, Nikhil Naik, Nitish Shirish Keskar, Namrata Anand, Raphael R Eguchi, Po-Ssu Huang, and Richard Socher. Progen: Language modeling for protein generation. arXiv preprint arXiv:2004.03497,

    2004 arXiv

  6. [2023]

    2021 fda approvals

    Asher Mullard. 2021 fda approvals. Nat. Rev. Drug Discov., 21(2):83–88,

    2021

Pith tools

Reviewed August 8, 2026 · model on record in the stance chip above.