Pith. sign in

REVIEW 2 major objections 4 minor 45 references

Smooth $\%$MinMax: A Differentiable Relaxation for Codon Harmonization

T0 review · 2 major / 4 minor · reviewed 2026-07-11 · grok-4.5

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

Pith's one-line read A smooth, differentiable form of the %MinMax codon-usage metric makes neural profile matching possible for codon harmonization.

desk verdict Clean, usable differentiable relaxation of %MinMax that does exactly what it claims in probability space; the only real gap is the missing discretization step. read the letter →

arxiv 2607.03881 v1 pith:J46Y7NSJ submitted 2026-07-04 q-bio.QM cs.LG

Yoonho Jeong , Hyunwoo Choi , Ryan Fernandez Medina Hariri , Eok Kyun Lee , Seung Seo Lee , Insung S. Choi This is my paper
classification q-bio.QMcs.LG
keywords codonharmonizationSmooth%MinMaxdifferentiablerelaxationsynonymousdesignneuralsequenceprofilematchingheterologousexpression
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

Codon harmonization tries to redesign a gene so a foreign host will translate it with the same local mix of common and rare codons that the original gene had, because those patterns can matter for how the protein folds while it is still being made. The usual %MinMax score that measures those local patterns only works on finished, discrete codon sequences, so it cannot be used directly with modern neural designers that update codon probabilities by gradient descent. This paper constructs Smooth %MinMax by replacing each hard codon choice with a probability-weighted average of synonymous-codon usages and replacing the hard positive/negative branch with a soft sigmoid gate. The new score stays close to the classical profile, keeps the same signed meaning (positive = enrichment of frequent codons, negative = enrichment of rare ones), and supplies usable gradients. In human-to-E. coli experiments the smooth profiles track the hard ones and the matching loss falls under gradient updates, showing that classical harmonization profiles can now sit inside neural sequence-design loops.

What carries the argument

Smooth %MinMax (%MinMax_[s]): the residue-level codon-usage value is replaced by its expectation under the synonymous-codon softmax, the window deviation is formed as usual, and the hard max/min denominator switch is replaced by a sigmoid-gated convex combination controlled by a sharpness hyperparameter β.

What would settle it

Take the optimized synonymous-codon probability distributions, sample or greedily decode discrete coding sequences, recompute classical hard %MinMax profiles on those sequences, and check whether they still match the human target profiles under the E. coli codon table; large systematic divergence would refute the claim that the smooth objective is a practical bridge.

Watch

Extended reading notes

Core claim

Smooth %MinMax, obtained by probability-weighted synonymous-codon usage and a sigmoid-gated interpolation between the positive and negative normalization branches, closely approximates conventional hard %MinMax while remaining fully differentiable with respect to codon probabilities and model parameters, thereby enabling gradient-based profile matching for neural codon harmonization.

Load-bearing premise

That matching profiles in soft probability space with a fixed smoothing strength will still produce useful discrete codon sequences once the probabilities are later turned into single codon choices.

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

2 major / 4 minor

Summary. The manuscript introduces Smooth %MinMax (%MinMax_[s]), a differentiable relaxation of the classical hard %MinMax (%MinMax_[h]) profile used in codon harmonization. Discrete codon-usage values are replaced by probability-weighted averages over synonymous codons, and the hard branch between positive and negative normalization denominators is replaced by a sigmoid-gated interpolation controlled by a sharpness parameter β. The authors show that, under one-hot synonymous distributions, %MinMax_[s] recovers %MinMax_[h] to high numerical fidelity (MAE, MSE, Pearson correlation versus β; Section 3.4, Figure 1) and that the smooth objective yields nonzero gradients that reduce profile-matching MAE from ~37 to ~7 in a simple BiLSTM model for human-to-E. coli harmonization (Section 4, Figure 2). Code and data are released.

Significance. If the technical claim holds, the work supplies a missing differentiable objective that lets classical profile-based codon harmonization be used inside gradient-based neural codon design. The formulation is transparent, the one-hot approximation study is quantitative, and the training curves confirm both loss decrease and gradient flow. Public code further strengthens reproducibility. The contribution is modest but useful: it does not claim new biological designs or expression gains, only a practical bridge between %MinMax profiles and probabilistic sequence models. That bridge is of clear interest to the synthetic-biology and computational-protein-design communities.

major comments (2)
  1. Section 4 and Conclusions: the central practical claim that %MinMax_[s] is a bridge to neural synonymous-sequence design is only partially supported. Optimization is performed entirely in probability space; the manuscript never samples discrete codon sequences from the optimized distributions, never recomputes hard %MinMax_[h] on those sequences, and never compares the resulting profiles (or any other metric) against CHARMING, MOSAIC, or a simple CAI baseline. Without this discretization step the utility for actual gene design remains untested, even though the mathematical differentiability claim itself is sound.
  2. Section 3.4: β = 0.1 is selected by a visual/MAE trade-off on one-hot human sequences and then fixed for all subsequent experiments. No sensitivity analysis is reported for the neural optimization itself (e.g., whether β = 0.01 or 1.0 changes final profile MAE or gradient norms). Because the gate sharpness directly affects both approximation fidelity and gradient magnitude near Δ̃_t = 0, a short ablation on the training objective would strengthen that the chosen value is not an accidental local optimum.
minor comments (4)
  1. Throughout: the notation %MinMax_[s] / %MinMax_[h] is clear, but the manuscript also uses several near-identical glyph variants (e.g., %MinMax, %MinMax[s]) that render inconsistently; a single consistent macro would improve readability.
  2. Figure 2a: the architecture diagram is helpful, yet the precise definition of the profile-context radius r (and why r = 5 was chosen after a pre-screen) is only stated in prose; adding the radius to the figure legend would make the input construction self-contained.
  3. Section 3.2: the usage frequency u(c) is defined as occurrences per 1 000 codons; a brief note on which codon-usage table (and organism version) was used for both human and E. coli would aid exact reproduction.
  4. Supporting Figure S1: six random test profiles are shown; reporting the mean test-set MAE (or Pearson) between smooth and hard profiles after training would quantify generalization more cleanly than visual inspection alone.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: Smooth %MinMax is an independent differentiable relaxation whose approximation and gradient utility are measured against the classical hard metric, not forced by construction or self-citation.

full rationale

The derivation is self-contained. Section 3.2 defines hard %MinMax_[h] from discrete codon usages, window deviations Δ_t, and a hard branch on the sign of Δ_t. Section 3.3 replaces the discrete usage with a probability-weighted average ñ_i and the hard branch with a sigmoid gate g_t = σ(β Δ̃_t), yielding a smooth profile M̃_t that is differentiable w.r.t. logits. When the synonymous distribution is one-hot, ñ_i recovers the hard usage exactly, so residual discrepancy is solely from the soft gate; the paper measures that discrepancy (MAE, MSE, Pearson) on real human sequences and selects β = 0.1 by explicit trade-off (Fig. 1). The subsequent neural experiment simply minimizes MAE between the host-side smooth profile and a fixed native target profile; loss decrease and nonzero gradients confirm that the objective is usable, but no fitted constant is later re-labeled a prediction, and no uniqueness theorem or load-bearing self-citation forces the result. The MOSAIC citation supplies only the source dataset. The acknowledged limitation (no discretization or wet-lab evaluation) is a scope gap, not circularity. Score 0 is therefore appropriate.

Assumptions & free parameters 3 free parameters · 4 assumptions · 1 invented entities

The central claim rests on the classical definition of %MinMax, standard codon-usage tables, the modeling choice that a soft probability distribution can stand in for a discrete codon, and a hand-chosen smoothing temperature. No new physical entities are postulated; the only invented object is the smooth metric itself.

free parameters (3)
  • β (sigmoid sharpness) = 0.1
    Fixed hyperparameter controlling the hard-branch approximation; selected by inspecting MAE/MSE/Pearson and visual profile agreement on one-hot human sequences (β = 0.1 chosen as compromise).
  • sliding-window size = 10
    Inherited from conventional %MinMax practice and fixed at 10 codons for all experiments.
  • profile-context radius r = 5
    How many neighboring target-profile values are fed to the prediction head; pre-screened and set to 5.
assumptions (4)
  • domain assumption Codon-usage frequencies u(c) measured per 1000 codons in a reference organism are the correct numerical basis for local rarity/commonness.
    Taken from the classical %MinMax definition (Rodriguez et al.) and used unchanged for both hard and smooth metrics.
  • ad hoc to paper A probability-weighted average of synonymous-codon usages is a valid continuous surrogate for the discrete codon-usage value.
    Core modeling step of Section 3.3; recovers the hard value exactly under one-hot distributions but is otherwise an unproven relaxation for optimization.
  • ad hoc to paper Sigmoid-gated interpolation between D+ and D− preserves the signed biological interpretation of %MinMax sufficiently for gradient-based design.
    Stated in Section 3.3–3.4; supported only by numerical approximation quality, not by biological validation.
  • domain assumption Matching a native %MinMax profile in the host codon-usage table is a meaningful objective for codon harmonization.
    Inherited from CHARMING/MOSAIC literature; the paper does not re-derive or re-validate this design principle.
invented entities (1)
  • Smooth %MinMax (%MinMax_[s])
    purpose: Provide a differentiable scalar profile that can be used as a loss for gradient-based synonymous-codon optimization while approximating classical hard %MinMax.
    Defined by the two relaxations in Section 3.3; no independent experimental handle outside the computational approximation and training curves shown in the paper.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Smooth $\%$MinMax: A Differentiable Relaxation for Codon Harmonization." pith.science (2026). https://pith.science/paper/J46Y7NSJ

@misc{pith2026260703881,
  author       = {Pith},
  title        = {Pith review of: Smooth $\%$MinMax: A Differentiable Relaxation for Codon Harmonization},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/J46Y7NSJ}},
  note         = {Machine review of arXiv:2607.03881}
}
abstract

Codon harmonization aims to adapt the coding sequences for heterologous expression while preserving the native-like patterns of frequent and rare codons that may influence local translation dynamics and co-translational protein folding. However, widely used harmonization metrics, such as $\%$MinMax, are defined on discrete codon sequences and are, therefore, not readily compatible with gradient-based neural codon design. Here, we introduce Smooth $\%$MinMax, denoted as $\%{\rm MinMax}_{(s)}$, a differentiable relaxation of the conventional hard $\%$MinMax metric, denoted as $\%{\rm MinMax}_{(h)}$. $\%{\rm MinMax}_{(s)}$ replaces the discrete codon-usage values with probability-weighted synonymous-codon usage values and replaces the hard $\%$Max/$\%$Min branch with a sigmoid-gated interpolation. This formulation preserves the signed interpretation of $\%{\rm MinMax}_{(h)}$, while enabling optimization with respect to the synonymous-codon probabilities and learnable parameters. In human-to-Escherichia coli codon harmonization experiments, $\%{\rm MinMax}_{(s)}$ closely approximates $\%{\rm MinMax}_{(h)}$ and supports gradient-based profile matching in synonymous-codon probability space. These results suggest $\%{\rm MinMax}_{(s)}$ as a practical bridge between profile-based codon harmonization and neural synonymous-sequence design.

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

45 extracted references · 1 canonical work pages

  1. [1]

    codon-usage bias

    Introduction The choice among synonymous codons is an important design consideration in the heterologous gene expression, in which a gene from one organism is expressed in a different host organism. Although synonymous codons encode the same amino acid, they are not used equally across species. Instead, each organism exhibits the characteristic patterns o...

  2. [2]

    Background & Related Work 2.1. Design Objectives and Evaluation Metrics in Codon Optimization Codon optimization studies have used a wide range of objectives and evaluation metrics to guide synonymous-sequence design. Among these metrics, CAI is one of the most widely used measures of host adaptation. CAI was originally developed to quantify how closely a...

  3. [3]

    Problem Setup for Synonymous-Codon Sequence Design Let 𝐴𝐴= (𝑛𝑛(1), … , 𝑛𝑛(𝑇𝑇)) denote an amino-acid sequence of length 𝑇𝑇

    𝐒𝐒𝐒𝐒𝐒𝐒𝐒𝐒𝐒𝐒𝐒𝐒 %𝑴𝑴𝑴𝑴𝑴𝑴𝑴𝑴𝑴𝑴𝑴𝑴 3.1. Problem Setup for Synonymous-Codon Sequence Design Let 𝐴𝐴= (𝑛𝑛(1), … , 𝑛𝑛(𝑇𝑇)) denote an amino-acid sequence of length 𝑇𝑇. For each amino acid 𝑛𝑛(𝑀𝑀) at position 𝑀𝑀, let 𝒮𝒮(𝑛𝑛(𝑀𝑀)) denote the set of synonymous codons that encode 𝑛𝑛(𝑀𝑀). The 5 problem of synonymous codon sequence design is to construct a coding sequence 𝒞𝒞= ...

  4. [4]

    Error bars indicate the standard deviation across 20,070 human coding sequences

    (b) MAE values between %𝑀𝑀𝑀𝑀𝑛𝑛𝑀𝑀𝑛𝑛𝑛𝑛[𝑠𝑠] and %𝑀𝑀𝑀𝑀𝑛𝑛𝑀𝑀𝑛𝑛𝑛𝑛[ℎ] profiles as a function of 𝛽𝛽. Error bars indicate the standard deviation across 20,070 human coding sequences. (c) Representative profile comparisons for the four values of 𝛽𝛽. The %𝑀𝑀𝑀𝑀𝑛𝑛𝑀𝑀𝑛𝑛𝑛𝑛[ℎ] profile is shown as a black dashed line, and the corresponding %𝑀𝑀𝑀𝑀𝑛𝑛𝑀𝑀𝑛𝑛𝑛𝑛[𝑠𝑠] profile is shown...

  5. [5]

    Experiments We evaluated whether %𝑀𝑀𝑀𝑀𝑛𝑛𝑀𝑀𝑛𝑛𝑛𝑛[𝑠𝑠] could serve as a differentiable objective for neural codon harmonization. For the proof-of-concept analysis, we used the human coding-sequence dataset previously employed in MOSAIC.18 The dataset comprised 20,070 protein-coding sequences derived from T2T-CHM13, a reference genome assembly of Homo sapiens....

  6. [6]

    Conclusions 13 In this study, we introduced 𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑡𝑡ℎ %𝑀𝑀𝑀𝑀𝑛𝑛𝑀𝑀𝑛𝑛𝑛𝑛, denoted %𝑀𝑀𝑀𝑀𝑛𝑛𝑀𝑀𝑛𝑛𝑛𝑛[𝑠𝑠], as a differentiable relaxation of the conventional hard %𝑀𝑀𝑀𝑀𝑛𝑛𝑀𝑀𝑛𝑛𝑛𝑛 metric, denoted %𝑀𝑀𝑀𝑀𝑛𝑛𝑀𝑀𝑛𝑛𝑛𝑛[ℎ], for profile- based neural codon harmonization. By replacing discrete codon-usage values with probability- weighted synonymous-codon usage values and replac...

  7. [7]

    Code Availability The source code and data used in this study are available at: https://github.com/CIS- group/SmoothMinMax

  8. [8]

    Codon bias and heterologous protein expression

    Gustafsson, C., Govindarajan, S., Minshull, J. Codon bias and heterologous protein expression. Trends Biotechnol. 2004, 22(7), 346-353

    2004

Show all 45 references
  1. [9]

    E., Claassens, N

    Quax, T. E., Claassens, N. J., Söll, D., Van Der Oost, J. Codon bias as a means to fine-tune gene expression. Mol. Cell 2015, 59(2), 149-161

    2015

  2. [10]

    T., Udayasuriyan, V ., Bhadana, V

    Parvathy, S. T., Udayasuriyan, V ., Bhadana, V . Codon usage bias. Mol. Biol. 2022, 49(1), 539-565

    2022

  3. [11]

    The selective advantage of synonymous codon usage bias in Salmonella

    Brandis, G., Hughes, D. The selective advantage of synonymous codon usage bias in Salmonella. PLoS Genet. 2016, 12(3), e1005926

    2016

  4. [12]

    Predicting synonymous codon usage and optimizing the heterologous gene for expression in E

    Tian, J., Yan, Y ., Yue, Q., Liu, X., Chu, X., Wu, N., Fan, Y . Predicting synonymous codon usage and optimizing the heterologous gene for expression in E. coli. Sci. Rep. 2017, 7(1), 9926

    2017

  5. [13]

    V ., 14 Komar, A

    Buhr, F., Jha, S., Thommen, M., Mittelstaet, J., Kutz, F., Schwalbe, H., Rodnina, M. V ., 14 Komar, A. A. Synonymous codons direct cotranslational folding toward different protein conformations. Mol. Cell 2016, 61(3), 341-351

    2016

  6. [14]

    A., Sharma, S., Sobott, F., Loenarz, C., Oppermann, U., Gileadi, O

    Burgess-Brown, N. A., Sharma, S., Sobott, F., Loenarz, C., Oppermann, U., Gileadi, O. Codon optimization can improve expression of human genes in Escherichia coli: A multi- gene study. Protein Expr. Purif. 2008, 59(1), 94-102

    2008

  7. [15]

    B., Zhang, X., Ma, F

    Al-Hawash, A. B., Zhang, X., Ma, F. Strategies of codon optimization for high-level heterologous protein expression in microbial expression systems. Gene Rep. 2017, 9, 46-53

    2017

  8. [16]

    G., Leung, K

    Gvritishvili, A. G., Leung, K. W., Tombran-Tink, J. Codon preference optimization increases heterologous PEDF expression. PLoS One 2010, 5(11), e15056

    2010

  9. [17]

    C., Chan, H

    Kwon, K. C., Chan, H. T., León, I. R., Williams-Carrier, R., Barkan, A., Daniell, H. Codon optimization to enhance expression yields insights into chloroplast translation. Plant Physiol. 2016, 172(1), 62-77

    2016

  10. [18]

    I., Suzuki, T

    Inouye, S., Sahara-Miura, Y ., Sato, J. I., Suzuki, T. Codon optimization of genes for efficient protein expression in mammalian cells by selection of only preferred human codons. Protein Expr. Purif. 2015, 109, 47-54

    2015

  11. [19]

    E., Claassens, N

    Quax, T. E., Claassens, N. J., Söll, D., van der Oost, J. Codon bias as a means to fine-tune gene expression. Mol. Cell 2015, 59(2), 149-161

    2015

  12. [20]

    F., Clark, P

    Clarke IV , T. F., Clark, P. L. Rare codons cluster. PloS One 2008, 3(10), e3412

    2008

  13. [21]

    H., Dang, Y ., Zhou, Z., Wu, C., Zhao, F., Sachs, M

    Yu, C. H., Dang, Y ., Zhou, Z., Wu, C., Zhao, F., Sachs, M. S., Liu, Y . Codon usage influences the local rate of translation elongation to regulate co-translational protein folding. Mol. Cell 2015, 59(5), 744-754

    2015

  14. [22]

    J., Kincaid, R

    Angov, E., Hillier, C. J., Kincaid, R. L., Lyon, J. A. Heterologous protein expression is enhanced by harmonizing the codon usage frequencies of the target gene with those of the expression host. PloS One 2008, 3(5), e2189

    2008

  15. [23]

    M., Mease, R

    Angov, E., Legler, P. M., Mease, R. M. Adjustment of codon usage frequencies by codon harmonization improves protein expression and folding. Methods Mol. Biol. 2011, 705, 1-13

    2011

  16. [24]

    J., Clark, P

    Wright, G., Rodriguez, A., Li, J., Milenkovic, T., Emrich, S. J., Clark, P. L. CHARMING: harmonizing synonymous codon usage to replicate a desired codon usage pattern. Protein Sci. 2022, 31(1), 221-231

    2022

  17. [25]

    Jeong, Y ., Yang, C., Hariri, R. F. M., Kim, J., Lee, E. K., Lee, Y ., Kim, W. J., Lee, S. S., Choi, I. S. M OSAIC: codon harmonization of Monte Carlo-based simulated annealing for linked codons in heterologous protein expression. ACS Synth. Biol. 2026, 15(6), 2293-2301

    2026

  18. [26]

    J., Siliakus, M

    Claassens, N. J., Siliakus, M. F., Spaans, S. K., Creutzburg, S. C., Nijsse, B., Schaap, P. J., Quax, T. E. F., Van Der Oost, J. Improving heterologous membrane protein production in Escherichia coli by combining transcriptional tuning and codon usage algorithms. PLoS One 2017...

    2017

  19. [27]

    S., Goegebeur, S., Delmulle, T., De Mol, M

    Willems, T., Hectors, W., Rombaut, J., De Rop, A. S., Goegebeur, S., Delmulle, T., De Mol, M. L., De Maeseneire, S. L., Soetaert, W. K. An exploratory in silico comparison of open- source codon harmonization tools. Microb. Cell Factories 2023, 22(1), 227

    2023

  20. [28]

    Rodriguez, A., Wright, G., Emrich, S., Clark, P. L. %MinMax: a versatile tool for calculating 15 and comparing synonymous codon usage and its impact on protein folding. Protein Sci. 2018, 27(1), 356-362

    2018

  21. [29]

    J., Lindner, A

    Fallahpour, A., Gureghian, V ., Filion, G. J., Lindner, A. B., Pandi, A. CodonTransformer: a multispecies codon optimizer using context-aware neural networks. Nat. Commun. 2025, 16(1), 3205

    2025

  22. [30]

    Codon optimization with deep learning to enhance protein expression

    Fu, H., Liang, Y ., Zhong, X., Pan, Z., Huang, L., Zhang, H., Xu, Y ., Zhou, W., Liu, Z. Codon optimization with deep learning to enhance protein expression. Sci. Rep. 2020, 10(1), 17617

    2020

  23. [31]

    ICOR: improving codon optimization with recurrent neural networks

    Jain, R., Jain, A., Mauro, E., LeShane, K., Densmore, D. ICOR: improving codon optimization with recurrent neural networks. BMC Bioinform. 2023, 24(1), 132

    2023

  24. [32]

    R., Yan, Y ., Agrawal, P., Waight, A

    Goulet, D. R., Yan, Y ., Agrawal, P., Waight, A. B., Mak, A. N. S., Zhu, Y . Codon optimization using a recurrent neural network. J. Comput. Biol. 2023, 30(1), 70-81

    2023

  25. [33]

    Ravi, S., Sharma, T., Yip, M., Yang, H., Xie, J., Gao, G., Tai, P. W. A deep learning model trained on expressed transcripts across different tissue types reveals cell-type codon- optimization preferences. Nucleic Acids Res. 2025, 53(6), gkaf233

    2025

  26. [34]

    CodonBERT: a BERT-based architecture tailored for codon optimization using the cross-attention mechanism

    Ren, Z., Jiang, L., Di, Y ., Zhang, D., Gong, J., Gong, J., Jiang, Q., Fu, Z., Sun, P., Zhou, B., Ni, M. CodonBERT: a BERT-based architecture tailored for codon optimization using the cross-attention mechanism. Bioinformatics 2024, 40(7), btae330

    2024

  27. [35]

    High-Codon: a deep learning-based codon optimization tool for enhanced heterologous protein expression in Escherichia coli

    Li, J., Dong, X., Liu, J. High-Codon: a deep learning-based codon optimization tool for enhanced heterologous protein expression in Escherichia coli. bioRxiv 2025, 656984. DOI: 10.1101/2025.05.30.656984

    2025 doi

  28. [36]

    M., Li, W

    Sharp, P. M., Li, W. H. The codon adaptation index-a measure of directional synonymous codon usage bias, and its potential applications. Nucleic Acids Res. 1987, 15(3), 1281-1295

    1987

  29. [37]

    C., Jahn, D

    Grote, A., Hiller, K., Scheer, M., Münch, R., Nörtemann, B., Hempel, D. C., Jahn, D. JCat: a novel tool to adapt codon usage of a target gene to its potential expression host. Nucleic Acids Res. 2005, 33(2), W526-W531

    2005

  30. [38]

    G., Garcia-Vallve, S

    Puigbò, P., Bravo, I. G., Garcia-Vallve, S. CAIcal: a combined set of tools to assess codon usage adaptation. Biol. Direct 2008, 3(1), 38

    2008

  31. [39]

    OPTIMIZER: a web server for optimizing the codon usage of DNA sequences

    Puigbo, P., Guzmán, E., Romeu, A., Garcia-Vallvé, S. OPTIMIZER: a web server for optimizing the codon usage of DNA sequences. Nucleic Acids Res. 2007, 35(2), W126- W131

    2007

  32. [40]

    A., Park, S

    Demissie, E. A., Park, S. Y ., Moon, J. H., Lee, D. Y . Comparative analysis of codon optimization tools: advancing toward a multi-criteria framework for synthetic gene design. J. Microbiol. Biotechnol. 2025, 35, e2411066

    2025

  33. [41]

    Şen, A., Kargar, K., Akgün, E., Pınar, M. Ç. Codon optimization: a mathematical programing approach. Bioinformatics 2020, 36(13), 4012-4020

    2020

  34. [42]

    H., Zhang, Y ., Huang, L

    Zhang, H., Zhang, L., Lin, A., Xu, C., Li, Z., Liu, K., Liu, B., Ma, X., Zhao, F., Jiang, H., Chen, C., Shen, H., Li, H., Mathews, D. H., Zhang, Y ., Huang, L. Algorithm for optimized mRNA design improves stability and immunogenicity. Nature 2023, 621(7978), 396-403

    2023

  35. [43]

    E., LaDue, M., Kumar, V ., Davis, W., Galda, A

    Alevras, D., Metkar, M., Friedhoff, T., Park, J. E., LaDue, M., Kumar, V ., Davis, W., Galda, A. Co-optimization of codon usage and mRNA secondary structure using quantum 16 computing. Proc. IEEE Int. Conf. Quantum Comput. Eng. (QCE) 2025, 1, 2346-2351

    2025

  36. [44]

    DeepCodon: a deep learning codon-optimization model to enhance protein expression

    Han, X., Shao, X., Shuo, L., Shi, Z., Huang, R., Chu, H., Zhang, H., Wang, R., Li, H., Liao, X., Cheng, J., Jiang, H. DeepCodon: a deep learning codon-optimization model to enhance protein expression. BioDesign Res. 2025, 100042

    2025

  37. [45]

    A new deep-learning-based approach for mRNA optimization: high Fidelity, computation efficiency, and multiple optimization factors

    Gong, Z., Jiang, Z., Gao, W., Zhuo, D., Ma, L. A new deep-learning-based approach for mRNA optimization: high Fidelity, computation efficiency, and multiple optimization factors. arXiv preprint 2025, arXiv:2505.23862. 17 SUPPORTING INFORMATION Figure S1. Six representative pro...

    2025 arXiv

Pith tools

Reviewed July 11, 2026 · model on record in the stance chip above.