Pith. sign in

REVIEW 3 major objections 4 minor 59 references

NbBench: Benchmarking Language Models for Comprehensive Nanobody Tasks

T0 review · 3 major / 4 minor · reviewed 2026-08-16 · deepseek-v4-flash

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

Pith's one-line read NbBench aims to be the first standardized nanobody benchmark, and its frozen evaluation shows no single language model dominates across tasks.

desk verdict A genuinely useful first nanobody benchmark whose model rankings are provisional until the authors fix an internal split-similarity contradiction, an affinity-task definition mix-up, and undisclosed pretraining overlap. read the letter →

arxiv 2505.02022 v2 pith:TY6LEGU2 submitted 2025-05-04 cs.LG q-bio.BM

Yiming Zhang , Koji Tsuda This is my paper · ORCID
classification cs.LGq-bio.BM
keywords NbBenchnanobodieslanguagemodelsbenchmarkbindingpredictionthermostabilitydevelopabilityrepresentationlearning
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

The paper introduces NbBench, a benchmark suite that aims to be the first standardized way to evaluate language models on nanobody tasks. It assembles eight tasks from nine datasets spanning structure annotation, binding prediction, and developability, and runs eleven pretrained protein, antibody, and nanobody language models through a frozen evaluation protocol. The central finding is that no single model wins across all tasks: antibody-specific models lead on antigen-binding and paratope predictions, structure-annotation tasks are effectively saturated, and regression tasks such as thermostability and affinity remain difficult for every model. The authors argue that NbBench matters because it replaces inconsistent single-task evaluations with a reproducible common protocol for nanobody modeling.

What carries the argument

The central object is NbBench itself: eight tasks across nine curated datasets, with clustering-based splits built by MMseqs2 at 70% sequence identity so that similar sequences stay in the same split. Evaluation is frozen: pretrained model weights stay fixed and only a task-specific MLP head is trained, which isolates the quality of the representations. For nanobody-antigen tasks, antigen embeddings are always produced by the same ESM-2 (650M) model and concatenated with the nanobody embedding, so differences between models reflect how well each nanobody encoder represents the variable domain. This shared pipeline is what lets the paper attribute observed performance gaps to the pretraining domain rather than to task formatting.

What would settle it

Compare each model's pretraining corpus against the test sequences of each NbBench task and remove any test sequence whose closest pretraining neighbor is above 70% identity, then rerun the eleven-model evaluation; if the reported task rankings change, the leakage controls are carrying the conclusion.

Watch

Extended reading notes

Core claim

NbBench claims to be the first comprehensive benchmark for nanobody modeling, and its evaluation of eleven frozen models maps out where current representations succeed and fail. Variable-region classification and nanobody-type classification are essentially solved at >98% accuracy. Antibody language models trained on large immunoglobulin corpora transfer best to antigen-related tasks, with AntiBERTa2-CSSP reaching 0.92 AUROC on hIL6 binding and 0.94 AUROC on paratope prediction. Regression on thermostability is modest at best (Spearman $\rho \approx 0.59$) and affinity prediction stays below 0.20 Spearman. The paper concludes that no universal model exists for nanobody tasks, and that future progress will require domain-matched pretraining combined with structure-aware or parameter-efficient fine-tuning.

Load-bearing premise

The load-bearing premise is that the clustering-based splits keep test nanobodies genuinely unseen, so that no model has already memorized its test examples during pretraining.

Editorial extensions

If this is right

  • Any new nanobody language model can be compared against eleven baselines on identical splits and metrics, so a reported improvement is attributable to the model rather than the evaluation protocol.
  • Structure-annotation tasks (VRCls and NbType) are saturated, so they should be treated as completed sanity checks rather than discriminative benchmarks.
  • For antigen-binding and paratope tasks, antibody-specific models such as AntiBERTa2 provide the strongest starting point among the tested frozen backbones.
  • Thermostability and affinity regressions are not solved by frozen sequence embeddings; the paper points to structure-aware or parameter-efficient fine-tuning as the next direction.
  • Because no model wins across all tasks, model selection for real nanobody engineering should be task-dependent.

Reading between the lines

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

  • Beyond the paper: the reported rankings depend on the leakage controls; if a model was pretrained on the same source database as its test set, its apparent edge could shrink once near-duplicate test sequences are removed.
  • Beyond the paper: fixing the antigen encoder to ESM-2 (650M) means binding-task results confound nanobody representations with that single antigen representation; ablating or varying the antigen encoder would show how much of the ranking is really about the nanobody model.
  • Beyond the paper: the paper's own limitation note implies that adding structure-aware tasks as nanobody structure databases grow is a natural next step; a plausible testable expectation is that paratope and binding rankings will shift once structural tokens are included.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 4 minor

Summary. NbBench is introduced as the first comprehensive benchmark suite for nanobody representation learning. The paper curates nine datasets into eight tasks spanning structure annotation (VRCls, CDRInf), binding prediction (SARS-CoV-2 and hIL6 antigen binding, paratope prediction), and biophysical properties (thermostability, polyreactivity, nanobody type, VHH affinity). Eleven pretrained protein, antibody, and nanobody language models are evaluated in a frozen setting with a fixed classifier head, using three random seeds. The main empirical findings are that antibody LMs dominate antigen-related tasks, regression tasks such as thermostability and affinity remain difficult for all models, and no single model wins across all tasks. The authors release code and data publicly.

Significance. If the benchmark construction and the comparison protocol are sound, NbBench would be a useful community resource: it standardizes task definitions, splits, and metrics for a domain that currently lacks a unified evaluation, and it provides a reproducible head-to-head comparison of eleven publicly available models. The paper also ships public code and datasets, which is a concrete strength. The findings about task difficulty and the absence of a universal winner are plausible and practically relevant. However, as detailed below, the validity of the comparison is currently undermined by an internal inconsistency in the redundancy control for the splits, by a contradictory description of the affinity task, and by unaddressed pretraining-overlap risks. These issues are load-bearing for the headline empirical claims and need to be fixed before the results can be trusted.

major comments (3)
  1. [Section 3.4, Appendix A.2, Section 3.1] The stated redundancy control is internally inconsistent. Section 3.4 says MMseqs2 clustering at a 70% sequence identity threshold assigns whole clusters to splits, which would imply that no cross-split sequence pair has identity above 70%. Yet Appendix A.2 reports that all tasks have average train-test similarities between 73% and 84%, and Section 3.1 says VRCls 'ensure[s] the average sequence similarity between training and test sets remains at 75%.' If the average pairwise identity exceeds 70%, then the 70% clustering threshold cannot have been enforced as described; conversely, if the clustering was enforced, the reported 73-84% figures must use a different definition of 'similarity' than the one implied by the clustering threshold. The paper must state exactly how the similarity was computed and report the actual distribution of cross-split pairwise identities. Because every ranking in Table 3 depends on the low-redundancy guarantee, this inconsistency needs to be resolved before the model comparison can be considered reliable.
  2. [Section 3.3, Table 1, Appendix A.2, Table 9] The affinity task definition is contradictory. Section 3.3 and Table 1 describe VHH Affinity Prediction as a regression task with the continuous score defined in Eq. (1), and Table 9 reports Spearman correlation, R2, RMSE, and MAE for Affinity-seq and Affinity-score. However, Appendix A.2 states: 'For affinity prediction, only sequences with binding scores ≥ 5 or ≤ 1 were retained for binary classification.' This describes a binary classification setup on a filtered subset, not the regression setup of Section 3.3. If the regression experiments were trained on the full score range, the appendix description is wrong; if a binary-filtered subset was used, then the regression metrics in Table 9 are not interpretable as stated, and the conclusion that 'regression remains hard' for affinity is not supported by the described data. Please clarify the exact filtering, the response variable used, and how Affinity-seq differs from Affinity-score.
  3. [Section 4, Table 2, Table 3] The fairness of the model comparison is not established because some baselines appear to have been pretrained on the same source databases used to construct the benchmark tasks. Specifically, NanoBERT was pretrained on INDI (Table 2) and is evaluated on VRCls/CDRInf tasks built from INDI, while VHHBERT was pretrained on VHHCorpus-2M and is evaluated on the AVIDa-SARS-CoV-2 dataset introduced in the same paper as VHHBERT. Since all models are frozen, a model that encountered test-distribution sequences during pretraining can produce better linear-probe results without actually displaying better transfer. The paper should quantify the sequence-level overlap between each task's test split and each model's pretraining corpus (for example, by an MMseqs2 search), or otherwise demonstrate that the rankings in Table 3 are not driven by pretraining contamination. Without this, the headline conclusion that 'antibody LMs excel on antigen-related tasks' is at risk.
minor comments (4)
  1. [Section 5.2] The text says the models were evaluated on 'the twelve tasks in NbBench,' but Table 1 lists eight tasks (with two binding datasets and two regression variants each). Please align the task count with the actual benchmark structure.
  2. [Appendix A.5] The text contains unresolved references: 'Table ?? provides a comprehensive overview' and 'Table 9, which has been reused for consistency.' Please update these to the correct table numbers and ensure that each table is cited where it is discussed.
  3. [Section 3.4, Table 1] The suffixes '-seq', '-tm', and '-score' are introduced as indicating sequence-clustering splits versus stratified splits, but the description is ambiguous for Affinity-seq and Affinity-score. Please state explicitly which split type applies to each of the four regression-variant tasks.
  4. [Appendix A.4] The definition of BLOSUM62 Recovery says 'scores are normalized if needed for comparability,' but the reported values in Table 7 are not accompanied by a statement of which normalization was applied. Please specify the normalization or say that raw BLOSUM62 scores are reported.

Circularity Check

2 steps flagged · score 3.0 of 10

Partial evaluation circularity: two baselines are scored on data from their own pretraining corpora; central conclusions otherwise independent.

  1. fitted input called prediction [Table 2 + Section 3.1 (VRCls/CDRInf)]
    "NanoBERT [26] is a RoBERTa-based model trained specifically on nanobody sequences to explore mutational space and thermostability. ... We collect data from the INDI database [14], using the abgenbank, manual, and structure subsets for training, and the patent subset for testing."

    Table 2 lists NanoBERT's pretraining corpus as 'INDI [14] (10M)', the same database that supplies VRCls and CDRInf, with the test split being the INDI patent subset. The paper only removes exact duplicates between benchmark train and test (Section 3.1), never sequences present in NanoBERT's pretraining corpus. For this model, VRCls/CDRInf accuracy is therefore partly a retrieval of pretraining inputs rather than a prediction on unseen nanobody sequences; the claimed generalization measurement reduces, for this row, to a fitted-input-called-prediction.

  2. other [Table 2 + Section 3.2 (SARS-CoV-2 binding)]
    "VHHBERT [15] is a RoBERTa-based model pre-trained on 2 million camelid nanobody sequences from VHHCorpus-2M. ... (1) SARS-CoV-2 [15]"

    The SARS-CoV-2 binding task is sourced from the same reference [15] (AVIDa-SARS-CoV-2) as VHHBERT's pretraining corpus (VHHCorpus-2M). NbBench reports no exclusion of the task VHH sequences from VHHBERT's pretraining, so VHHBERT's AUROC on this task can reflect memorization of pretraining data rather than transfer. At minimum, the benchmark design permits this particular prediction to reduce to pretraining lookup, so it is not demonstrated to be an independent generalization measurement.

full rationale

NbBench is an empirical benchmark rather than a derivation chain, and most of its headline conclusions (antibody LMs excel on antigen-related tasks, regression remains hard, no single model dominates) are read directly from Table 3 and are not implied by the benchmark construction. The Affinity task's use of MOQA [56] is an external experimental resource even though one NbBench author co-authored it; that is a provenance concern, not a circular derivation. The self-citation [30] in Related Work is not load-bearing. However, partial evaluation circularity exists for two nanobody-specific baselines: NanoBERT is pretrained on INDI and evaluated on INDI-derived VRCls/CDRInf, and VHHBERT is pretrained on VHHCorpus-2M [15] while the SARS-CoV-2 binding task is sourced from the same [15] resource, with no decontamination against pretraining corpora reported. For those rows, the reported generalization can reduce to recognition of pretraining inputs. The separate internal inconsistency between the claimed MMseqs2 70% clustering threshold (Section 3.4) and the reported 73-84% average train-test similarity (Appendix A.2) is a validity defect in the split protocol, not circularity. Because the dominant conclusions are not carried by the two affected model rows, the overall circularity is partial rather than total.

Assumptions & free parameters 4 free parameters · 4 assumptions · 0 invented entities

The central claim rests on benchmark design choices such as the clustering threshold, test sampling fraction, contact cutoff, and affinity thresholds, plus domain assumptions about frozen embeddings and split integrity. The affinity threshold entry is flagged because the paper internally contradicts itself on whether the task is regression or binary classification.

free parameters (4)
  • MMseqs2 sequence identity threshold = 0.7
    Chosen threshold for clustering-based train/test splits in Section 3.4; directly affects split difficulty and all reported results.
  • Test-set sampling fraction from low-frequency clusters = 0.2
    20% of low-frequency clusters are assigned to the test set, a hand-chosen value in Section 3.4 that affects generalization estimates.
  • Contact cutoff for paratope labeling = 5 Angstrom
    Residue pairs within 5 Angstrom in SAbDab-Nano complexes are labeled as interacting, a binary labeling choice without sensitivity analysis.
  • Affinity score thresholds = score >= 5 or <= 1
    Appendix A.2 states these thresholds for a binary classification version of the affinity task, conflicting with the regression definition in Section 3.3.
assumptions (4)
  • domain assumption Frozen language model embeddings, with only a lightweight MLP head trained, are sufficient to meaningfully benchmark nanobody representation quality.
    This is the core protocol in Section 4 and Section 5.1; if fine-tuning or structural features were required, the benchmark would not measure what it claims.
  • domain assumption MMseqs2 clustering at 70% identity keeps training and test sets sufficiently dissimilar to prevent leakage.
    Stated in Section 3.4; the paper's own appendix reports 73-84% average train-test similarity, which weakens this assumption.
  • domain assumption Models pretrained on the same source databases as the task data receive no unfair advantage.
    Implied by the fair comparison claim in Section 4; no pretraining-to-task overlap audit is performed.
  • domain assumption The affinity ratio in Eq. 1 is a valid continuous target for regression.
    VHH affinity is defined as a ratio of frequencies in Section 3.3; if this ratio is noisy or bimodal, the regression evaluation is not meaningful.

how reviews work

0 comments
Cite this review

Pith. "Pith review of NbBench: Benchmarking Language Models for Comprehensive Nanobody Tasks." pith.science (2026). https://pith.science/paper/TY6LEGU2

@misc{pith2026250502022,
  author       = {Pith},
  title        = {Pith review of: NbBench: Benchmarking Language Models for Comprehensive Nanobody Tasks},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/TY6LEGU2}},
  note         = {Machine review of arXiv:2505.02022}
}
read the original abstract

Nanobodies -- single-domain antibody fragments derived from camelid heavy-chain-only antibodies -- exhibit unique advantages such as compact size, high stability, and strong binding affinity, making them valuable tools in therapeutics and diagnostics. While recent advances in pretrained protein and antibody language models (PPLMs and PALMs) have greatly enhanced biomolecular understanding, nanobody-specific modeling remains underexplored and lacks a unified benchmark. To address this gap, we introduce NbBench, the first comprehensive benchmark suite for nanobody representation learning. Spanning eight biologically meaningful tasks across nine curated datasets, NbBench encompasses structure annotation, binding prediction, and developability assessment. We systematically evaluate eleven representative models -- including general-purpose protein LMs, antibody-specific LMs, and nanobody-specific LMs -- in a frozen setting. Our analysis reveals that antibody language models excel in antigen-related tasks, while performance on regression tasks such as thermostability and affinity remains challenging across all models. Notably, no single model consistently outperforms others across all tasks. By standardizing datasets, task definitions, and evaluation protocols, NbBench offers a reproducible foundation for assessing and advancing nanobody modeling.

Figures

Figures reproduced from arXiv: 2505.02022 by the authors.

Figure 1
Figure 1. Overview of NbBench. (a) Eight tasks spanning token- and sequence-level classification and regression. (b) Data sources include INDI, SAbDab, PLAbDab-nano, and task-specific datasets. (c) 10+ baselines from pretrained protein, antibody, and nanobody language models. (d) Visual representations of tasks grouped into structure understanding, binding, and biophysical and functional property prediction. (e) Standard eval… view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

59 extracted references · 49 canonical work pages

  1. [1]

    Development of therapeutic antibodies for the treatment of diseases

    Ruei-Min Lu, Yu-Chyi Hwang, I-Ju Liu, Chi-Chiu Lee, Han-Zen Tsai, Hsin-Jung Li, and Han-Chung Wu. Development of therapeutic antibodies for the treatment of diseases. Journal of biomedical science, 27:1–30, 2020

    work page 2020

  2. [2]

    Development trends for human monoclonal antibody therapeutics

    Aaron L Nelson, Eugen Dhimolea, and Janice M Reichert. Development trends for human monoclonal antibody therapeutics. Nature reviews drug discovery, 9(10):767–774, 2010

    work page 2010

  3. [3]

    Structure and function of immunoglobulins

    Harry W Schroeder Jr and Lisa Cavacini. Structure and function of immunoglobulins. Journal of allergy and clinical immunology, 125(2):S41–S52, 2010

    work page 2010

  4. [4]

    Nanobody engineering: computational modelling and design for biomedical and therapeutic applications

    Nehad S El Salamouni, Jordan H Cater, Lisanne M Spenkelink, and Haibo Yu. Nanobody engineering: computational modelling and design for biomedical and therapeutic applications. FEBS Open Bio, 15(2):236–253, 2025

    work page 2025

  5. [5]

    Nanobodies: natural single-domain antibodies

    Serge Muyldermans. Nanobodies: natural single-domain antibodies. Annual review of biochem- istry, 82(1):775–797, 2013. 9

    work page 2013

  6. [6]

    The therapeutic potential of nanobodies

    Ivana Jovˇcevska and Serge Muyldermans. The therapeutic potential of nanobodies. BioDrugs, 34(1):11–26, 2020

    work page 2020

  7. [7]

    Nanobody-based products as research and diagnostic tools

    Thomas De Meyer, Serge Muyldermans, and Anna Depicker. Nanobody-based products as research and diagnostic tools. Trends in biotechnology, 32 5:263–70, 2014

    work page 2014

  8. [8]

    Nanobodies as versatile tools to understand, diagnose, visualize and treat cancer

    Isabel Van Audenhove and Jan Gettemans. Nanobodies as versatile tools to understand, diagnose, visualize and treat cancer. EBioMedicine, 8:40–48, 2016

    work page 2016

Show all 59 references
  1. [9]

    Discovery of nanobodies: a comprehensive review of their applications and potential over the past five years

    Elena Alexander and Kam W Leong. Discovery of nanobodies: a comprehensive review of their applications and potential over the past five years. Journal of Nanobiotechnology, 22(1):661, 2024

    2024

  2. [10]

    Nanobody: a promising toolkit for molecular imaging and disease therapy

    Guangfa Bao, Ming Tang, Jun Zhao, and Xiaohua Zhu. Nanobody: a promising toolkit for molecular imaging and disease therapy. EJNMMI research, 11:1–13, 2021

    2021

  3. [11]

    Nanobodies as therapeutics: big opportunities for small antibodies

    Sophie Steeland, Roosmarijn Vandenbroucke, and Claude Libert. Nanobodies as therapeutics: big opportunities for small antibodies. Drug discovery today, 21(7):1076–1113, 2016

    2016

  4. [12]

    Observed antibody space: a resource for data mining next-generation sequencing of antibody repertoires

    Aleksandr Kovaltsuk, Jinwoo Leem, Sebastian Kelm, James Snowden, Charlotte M Deane, and Konrad Krawczyk. Observed antibody space: a resource for data mining next-generation sequencing of antibody repertoires. The Journal of Immunology, 201(8):2502–2509, 2018

    2018

  5. [13]

    Observed antibody space: A diverse database of cleaned, annotated, and translated unpaired and paired antibody sequences

    Tobias H Olsen, Fergus Boyles, and Charlotte M Deane. Observed antibody space: A diverse database of cleaned, annotated, and translated unpaired and paired antibody sequences. Protein Science, 31(1):141–146, 2022

    2022

  6. [14]

    Indi—integrated nanobody database for immunoinformatics

    Piotr Deszy´nski, Jakub Młokosiewicz, Adam V olanakis, Igor Jaszczyszyn, Natalie Castellana, Stefano Bonissone, Rajkumar Ganesan, and Konrad Krawczyk. Indi—integrated nanobody database for immunoinformatics. Nucleic Acids Research, 50(D1):D1273–D1281, 2022

    2022

  7. [15]

    A SARS-CoV-2 interaction dataset and VHH sequence corpus for antibody language models

    Hirofumi Tsuruta, Hiroyuki Yamazaki, Ryota Maeda, Ryotaro Tamura, and Akihiro Imura. A SARS-CoV-2 interaction dataset and VHH sequence corpus for antibody language models. In Advances in Neural Information Processing Systems 37, 2024

    2024

  8. [16]

    Evolutionary-scale prediction of atomic-level protein structure with a language model

    Zeming Lin, Halil Akin, Roshan Rao, Brian Hie, Zhongkai Zhu, Wenting Lu, Nikita Smetanin, Robert Verkuil, Ori Kabeli, Yaniv Shmueli, et al. Evolutionary-scale prediction of atomic-level protein structure with a language model. Science, 379(6637):1123–1130, 2023

    2023

  9. [17]

    Prottrans: Toward understanding the language of life through self-supervised learning

    Ahmed Elnaggar, Michael Heinzinger, Christian Dallago, Ghalia Rehawi, Yu Wang, Llion Jones, Tom Gibbs, Tamas Feher, Christoph Angerer, Martin Steinegger, et al. Prottrans: Toward understanding the language of life through self-supervised learning. IEEE transactions on pattern ...

    2021

  10. [18]

    Msa transformer

    Roshan M Rao, Jason Liu, Robert Verkuil, Joshua Meier, John Canny, Pieter Abbeel, Tom Sercu, and Alexander Rives. Msa transformer. In International Conference on Machine Learning, pages 8844–8856. PMLR, 2021

    2021

  11. [19]

    Evaluating protein transfer learning with tape

    Roshan Rao, Nicholas Bhattacharya, Neil Thomas, Yan Duan, Peter Chen, John Canny, Pieter Abbeel, and Yun Song. Evaluating protein transfer learning with tape. Advances in neural information processing systems, 32, 2019

    2019

  12. [20]

    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, 2020

    2004 arXiv

  13. [21]

    Language models enable zero-shot prediction of the effects of mutations on protein function

    Joshua Meier, Roshan Rao, Robert Verkuil, Jason Liu, Tom Sercu, and Alex Rives. Language models enable zero-shot prediction of the effects of mutations on protein function. Advances in neural information processing systems, 34:29287–29303, 2021

    2021

  14. [22]

    Deciphering antibody affinity maturation with language models and weakly supervised learning

    Jeffrey A Ruffolo, Jeffrey J Gray, and Jeremias Sulam. Deciphering antibody affinity maturation with language models and weakly supervised learning. arXiv preprint arXiv:2112.07782, 2021

    2021 arXiv

  15. [23]

    Ablang: an antibody language model for completing antibody sequences

    Tobias H Olsen, Iain H Moal, and Charlotte M Deane. Ablang: an antibody language model for completing antibody sequences. Bioinformatics Advances, 2(1):vbac046, 2022. 10

    2022

  16. [24]

    Enhancing antibody language models with structural information

    Justin Barton, Jacob D Galson, and Jinwoo Leem. Enhancing antibody language models with structural information. BioRxiv, pages 2023–12, 2024

    2023

  17. [25]

    Large scale paired antibody language models

    Henry Kenlay, Frédéric A Dreyer, Aleksandr Kovaltsuk, Dom Miketa, Douglas Pires, and Charlotte M Deane. Large scale paired antibody language models. PLOS Computational Biology, 20(12):e1012646, 2024

    2024

  18. [26]

    nanobert: a deep learning model for gene agnostic navigation of the nanobody mutational space

    Johannes Thorling Hadsund, Tadeusz Satława, Bartosz Janusz, Lu Shan, Li Zhou, Richard Röttger, and Konrad Krawczyk. nanobert: a deep learning model for gene agnostic navigation of the nanobody mutational space. Bioinformatics Advances, 4(1):vbae033, 2024

    2024

  19. [27]

    Nanoberta-asp: predicting nanobody paratope based on a pretrained roberta model

    Shangru Li, Xiangpeng Meng, Rui Li, Bingding Huang, and Xin Wang. Nanoberta-asp: predicting nanobody paratope based on a pretrained roberta model. BMC bioinformatics, 25(1):122, 2024

    2024

  20. [28]

    Nanobody–antigen interaction prediction with ensemble deep learning and prompt-based protein language models

    Juntao Deng, Miao Gu, Pengyan Zhang, Mingyu Dong, Tao Liu, Yabin Zhang, and Min Liu. Nanobody–antigen interaction prediction with ensemble deep learning and prompt-based protein language models. Nature Machine Intelligence, 6(12):1594–1604, 2024

    2024

  21. [29]

    Structure-aware protein self- supervised learning

    Can Chen, Jingbo Zhou, Fan Wang, Xue Liu, and Dejing Dou. Structure-aware protein self- supervised learning. Bioinformatics, 39(4):btad189, 2023

    2023

  22. [30]

    Hyena architecture enables fast and efficient protein language modeling

    Yiming Zhang, Bian Bian, and Manabu Okumura. Hyena architecture enables fast and efficient protein language modeling. iMetaOmics, 2(1):e45, 2025

    2025

  23. [31]

    Deciphering the language of antibodies using self-supervised learning

    Jinwoo Leem, Laura S Mitchell, James HR Farmery, Justin Barton, and Jacob D Galson. Deciphering the language of antibodies using self-supervised learning. Patterns, 3(7), 2022

    2022

  24. [32]

    Rapid discovery of high-affinity antibodies via massively parallel sequencing, ribosome display and affinity screening

    Benjamin T Porebski, Matthew Balmforth, Gareth Browne, Aidan Riley, Kiarash Jamali, Maximillian JLJ Fürst, Mirko Velic, Andrew Buchanan, Ralph Minter, Tristan Vaughan, et al. Rapid discovery of high-affinity antibodies via massively parallel sequencing, ribosome display and af...

    2024

  25. [33]

    On pre-trained language models for antibody

    Danqing Wang, Fei Ye, and Hao Zhou. On pre-trained language models for antibody. arXiv preprint arXiv:2301.12112, 2023

    2023 arXiv

  26. [34]

    Nanobody: outstanding features for diagnostic and therapeutic applications

    J-Pablo Salvador, Lluïsa Vilaplana, and M-Pilar Marco. Nanobody: outstanding features for diagnostic and therapeutic applications. Analytical and bioanalytical chemistry, 411:1703–1713, 2019

    2019

  27. [35]

    Unveiling the new chapter in nanobody engineering: advances in traditional construction and ai-driven optimization

    Jiwei Liu, Lei Wu, Anqi Xie, Weici Liu, Zhao He, Yuan Wan, and Wenjun Mao. Unveiling the new chapter in nanobody engineering: advances in traditional construction and ai-driven optimization. Journal of Nanobiotechnology, 23:87, 2025

    2025

  28. [36]

    Antibody design using deep learning: from sequence and structure design to affinity maturation

    Sara Joubbi, Alessio Micheli, Paolo Milazzo, Giuseppe Maccari, Giorgio Ciano, Dario Carda- mone, and Duccio Medini. Antibody design using deep learning: from sequence and structure design to affinity maturation. Briefings in Bioinformatics, 25(4):bbae307, 2024

    2024

  29. [37]

    Highly accurate protein structure prediction with alphafold

    John Jumper, Richard Evans, Alexander Pritzel, Tim Green, Michael Figurnov, Olaf Ron- neberger, Kathryn Tunyasuvunakool, Russ Bates, Augustin Žídek, Anna Potapenko, et al. Highly accurate protein structure prediction with alphafold. nature, 596(7873):583–589, 2021

    2021

  30. [38]

    High-resolution de novo structure prediction from primary sequence

    Ruidong Wu, Fan Ding, Rui Wang, Rui Shen, Xiwen Zhang, Shitong Luo, Chenpeng Su, Zuofan Wu, Qi Xie, Bonnie Berger, et al. High-resolution de novo structure prediction from primary sequence. BioRxiv, pages 2022–07, 2022

    2022

  31. [39]

    Structural modeling of nanobodies: a benchmark of state-of-the-art artificial intelli- gence programs

    Mario S Valdés-Tresanco, Mario E Valdés-Tresanco, Daiver E Jiménez-Gutiérrez, and Ernesto Moreno. Structural modeling of nanobodies: a benchmark of state-of-the-art artificial intelli- gence programs. Molecules, 28(10):3991, 2023

    2023

  32. [40]

    Peer: a comprehensive and multi-task benchmark for protein sequence understanding

    Minghao Xu, Zuobai Zhang, Jiarui Lu, Zhaocheng Zhu, Yangtian Zhang, Ma Chang, Runcheng Liu, and Jian Tang. Peer: a comprehensive and multi-task benchmark for protein sequence understanding. Advances in Neural Information Processing Systems, 35:35156–35173, 2022. 11

    2022

  33. [41]

    Flip: Benchmark tasks in fitness landscape inference for proteins

    Christian Dallago, Jody Mou, Kadina E Johnston, Bruce J Wittmann, Nicholas Bhattacharya, Samuel Goldman, Ali Madani, and Kevin K Yang. Flip: Benchmark tasks in fitness landscape inference for proteins. bioRxiv, pages 2021–11, 2021

    2021

  34. [42]

    Optimization of therapeutic antibodies by predicting antigen specificity from antibody sequence via deep learning

    Derek M Mason, Simon Friedensohn, Cédric R Weber, Christian Jordi, Bastian Wagner, Si- mon M Meng, Roy A Ehling, Lucia Bonati, Jan Dahinden, Pablo Gainza, et al. Optimization of therapeutic antibodies by predicting antigen specificity from antibody sequence via deep learning. ...

    2021

  35. [43]

    Para- pred: antibody paratope prediction using convolutional and recurrent neural networks

    Edgar Liberis, Petar Veliˇckovi´c, Pietro Sormanni, Michele Vendruscolo, and Pietro Liò. Para- pred: antibody paratope prediction using convolutional and recurrent neural networks. Bioinfor- matics, 34(17):2944–2950, 2018

    2018

  36. [44]

    Differences in the composition of the human antibody repertoire by b cell subsets in the blood

    Eva Szymanska Mroczek, Gregory C Ippolito, Tobias Rogosch, Kam Hon Hoi, Tracy A Hwangpo, Marsha G Brand, Yingxin Zhuang, Cun Ren Liu, David A Schneider, Michael Zemlin, et al. Differences in the composition of the human antibody repertoire by b cell subsets in the blood. Front...

    2014

  37. [45]

    Cov-abdab: the coronavirus antibody database

    Matthew IJ Raybould, Aleksandr Kovaltsuk, Claire Marks, and Charlotte M Deane. Cov-abdab: the coronavirus antibody database. Bioinformatics, 37(5):734–735, 2021

    2021

  38. [46]

    Disease diagnostics using machine learning of immune receptors

    Maxim E Zaslavsky, Erin Craig, Jackson K Michuda, Nidhi Sehgal, Nikhil Ram-Mohan, Ji- Yeun Lee, Khoa D Nguyen, Ramona A Hoh, Tho D Pham, Katharina Röltgen, et al. Disease diagnostics using machine learning of immune receptors. Biorxiv, pages 2022–04, 2022

    2022

  39. [47]

    Critical contribution of vh–vl interaction to reshaping of an antibody: The case of humanization of anti-lysozyme antibody, hyhel-10

    Takeshi Nakanishi, Kouhei Tsumoto, Akiko Yokota, Hidemasa Kondo, and Izumi Kumagai. Critical contribution of vh–vl interaction to reshaping of an antibody: The case of humanization of anti-lysozyme antibody, hyhel-10. Protein Science, 17(2):261–270, 2008

  40. [48]

    Wei, Zelda Mariet, Poomarin Phloyphisut, Hidetoshi Shimokawa, Joseph R

    Hirofumi Tsuruta, Hiroyuki Yamazaki, Ryota Maeda, Ryotaro Tamura, Jennifer N. Wei, Zelda Mariet, Poomarin Phloyphisut, Hidetoshi Shimokawa, Joseph R. Ledsam, Lucy Colwell, and Akihiro Imura. A VIDa-hIL6: A large-scale VHH dataset produced from an immunized alpaca for predictin...

    2023

  41. [49]

    Nbthermo: a new thermostability database for nanobodies

    Mario S Valdés-Tresanco, Mario E Valdés-Tresanco, Esteban Molina-Abad, and Ernesto Moreno. Nbthermo: a new thermostability database for nanobodies. Database, 2023:baad021, 2023

    2023

  42. [50]

    Tempro: nanobody melting temperature estima- tion model using protein embeddings

    Jerome Anthony E Alvarez and Scott N Dean. Tempro: nanobody melting temperature estima- tion model using protein embeddings. Scientific Reports, 14(1):19074, 2024

    2024

  43. [51]

    Prediction of protein biophysical traits from limited data: a case study on nanobody thermostability through nanomelt

    Aubin Ramon, Mingyang Ni, Olga Predeina, Rebecca Gaffey, Patrick Kunz, Shimobi Onuoha, and Pietro Sormanni. Prediction of protein biophysical traits from limited data: a case study on nanobody thermostability through nanomelt. In mAbs, volume 17, page 2442750. Taylor & Francis, 2025

    2025

  44. [52]

    An in silico method to assess antibody fragment polyreactivity

    Edward P Harvey, Jung-Eun Shin, Meredith A Skiba, Genevieve R Nemeth, Joseph D Hurley, Alon Wellner, Ada Y Shaw, Victor G Miranda, Joseph K Min, Chang C Liu, et al. An in silico method to assess antibody fragment polyreactivity. Nature communications, 13(1):7554, 2022

    2022

  45. [53]

    Gp-nano: a geometric graph network for nanobody polyreactivity prediction

    Xiaolong Zhou, Qianmu Yuan, Shuangjia Zheng, Yu Wang, and Yuedong Yang. Gp-nano: a geometric graph network for nanobody polyreactivity prediction. In 2024 IEEE International Conference on Bioinformatics and Biomedicine (BIBM), pages 161–167. IEEE, 2024

    2024

  46. [54]

    Plabdab-nano: a database of camelid and shark nanobodies from patents and literature

    Gemma L Gordon, Alexander Greenshields-Watson, Parth Agarwal, Ashley Wong, Fergus Boyles, Alissa Hummer, Ana G Lujan Hernandez, and Charlotte M Deane. Plabdab-nano: a database of camelid and shark nanobodies from patents and literature. Nucleic Acids Research, 53(D1):D535–D542, 2025

    2025

  47. [55]

    Assessing antibody and nanobody nativeness for hit selection and humanization with abnativ

    Aubin Ramon, Montader Ali, Misha Atkinson, Alessio Saturnino, Kieran Didi, Cristina Visentin, Stefano Ricagno, Xing Xu, Matthew Greenig, and Pietro Sormanni. Assessing antibody and nanobody nativeness for hit selection and humanization with abnativ. Nature Machine Intelligence...

    2024

  48. [56]

    Extensive antibody search with whole spectrum black-box optimization

    Andrejs Tuˇcs, Tomoyuki Ito, Yoichi Kurumida, Sakiya Kawada, Hikaru Nakazawa, Yutaka Saito, Mitsuo Umetsu, and Koji Tsuda. Extensive antibody search with whole spectrum black-box optimization. Scientific Reports, 14(1):552, 2024

    2024

  49. [57]

    Mmseqs2 enables sensitive protein sequence searching for the analysis of massive data sets

    Martin Steinegger and Johannes Söding. Mmseqs2 enables sensitive protein sequence searching for the analysis of massive data sets. Nature biotechnology, 35(11):1026–1028, 2017

    2017

  50. [58]

    sdab-db: the single domain antibody database, 2018

    Emily E Wilton, Michael P Opyr, Senthilkumar Kailasam, Ronja F Kothe, and Hans-Joachim Wieden. sdab-db: the single domain antibody database, 2018

    2018

  51. [59]

    easy-search

    Baris E Suzek, Hongzhan Huang, Peter McGarvey, Raja Mazumder, and Cathy H Wu. Uniref: comprehensive and non-redundant uniprot reference clusters. Bioinformatics, 23(10):1282– 1288, 2007. A Appendix A.1 Experimental Settings for Tasks All tasks in NbBench are trained using a sh...

    2007

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.