A preregistered pilot to validate short peptide myostatin inhibitors for muscle preservation and recovery

Track 1: Muscle: Performance, Preservation, and Recovery

1. Abstract

Myostatin (GDF-8) is a validated negative regulator of skeletal muscle growth and a direct fit for the muscle track of this call. This proposal tests whether a computationally prioritized short peptide can show direct, reproducible binding to mature human myostatin in a focused biophysical pilot. The study is intentionally narrow: one target, one primary assay, and a 3-peptide panel consisting of a literature-derived reference scaffold, one computationally optimized lead, and one preregistered negative control peptide. The principal investigator is Pavel Pravdin, a health-tech founder and machine-learning engineer with an ORCID-linked publication record in Scientific Reports, IEEE EMBC, and Sensors, working at the intersection of engineering, computer science, and applied biomedical research. The project uses commercial peptide synthesis, recombinant protein procurement, and surface plasmon resonance as the core validation workflow, followed by a limited counter-screen panel using GDF-11 as the closest-family cross-reactivity check and activin A as a broader selectivity comparator. Any candidate that clears the primary binding gate will then undergo a minimal serum or plasma stability assessment. No human tissue, stem cells, animals, or in vivo procedures are included. The pilot is structured as a 6-month project, with the decisive experimental work completed in the first 8 weeks. All raw assay outputs, fitted parameters, and methods will be shared openly regardless of outcome.

2. Introduction

Myostatin is one of the clearest molecular brakes on skeletal muscle growth and is directly relevant to muscle preservation, recovery, and lean-mass maintenance [1,2]. While myostatin has been pursued with antibodies and other larger biologics, short linear peptide inhibitors remain comparatively underexplored despite a structurally defined binding interface and clear antagonist precedents in the myostatin pathway [3]. This makes myostatin a good fit for a focused pilot in next-generation human enhancement research.

Short linear peptides were chosen for this pilot because they are fast to synthesize, inexpensive to iterate, and well suited to an open validation workflow. Their known limitations include proteolytic instability, weaker affinity, and poor permeability relative to larger modalities.

The practical question is not whether myostatin matters biologically. That is already well established. The practical question is whether a short peptide nominated by a computational discovery workflow can show real, reproducible target engagement under matched biophysical conditions. That is the right first gate before any larger functional program. Without clean target-engagement data, downstream functional assays are harder to interpret and more expensive to scale.

The primary objective of this study is a preregistered target-engagement and workflow-validation pilot, not a claim of therapeutic readiness from binding alone.

This proposal starts from a mature computational position rather than an open-ended discovery effort. The target structure, literature scaffold, candidate-generation workflow, and initial ranking have already been completed in-house. The requested funding is therefore not for exploratory target hunting. It is for one preregistered validation cycle on the most mature muscle-track program currently available in our pipeline.

The project is also explicitly interdisciplinary. The PI is a machine-learning engineer and founder applying computational modeling, workflow design, and external wet-lab execution to a biology problem with translational relevance. That profile matches the call’s emphasis on engineering and computer science integrated with life-science questions.

Experimental execution in this pilot will be carried out through commercial synthesis and contracted assay providers, while the PI will manage computational prioritization, preregistration, vendor coordination, and data interpretation.

Hypotheses

H1. Primary hypothesis

At least one computationally selected short peptide candidate will show measurable direct binding to recombinant mature human myostatin in a surface plasmon resonance assay.

H2. Secondary hypothesis

The top candidate will bind at least comparably to a literature-derived myostatin reference peptide tested under the same assay conditions.

H3. Secondary hypothesis

The top candidate will show stronger engagement of myostatin than activin A under matched conditions, while the GDF-11 result will be interpreted as a closest-family cross-reactivity readout rather than a stand-alone selectivity threshold.

3. Methods

Participants/Sample

There are no human or animal participants in this study. The experimental sample consists of:

three synthesized peptides
recombinant mature human myostatin
recombinant GDF-11 as the closest-family cross-reactivity comparator
recombinant activin A as a broader selectivity comparator

The 3-peptide panel is the minimum interpretable design for this pilot: one literature-derived reference scaffold, one computationally optimized lead, and one preregistered negative control peptide. The negative control will be a matched-composition scrambled variant designed to preserve length and broad physicochemical properties while disrupting the expected binding pattern, with the exact sequence and randomization rule locked in the public preregistration before synthesis. This is sufficient to answer the central question of whether an already-ranked candidate set can outperform or match a literature reference under matched assay conditions.

The study is designed as a pilot and is not intended to support population-level inference. Replicate assay runs are included to evaluate reproducibility rather than to claim large-sample statistical certainty.

Materials and Procedures

Computational locking phase

Before any wet-lab spending, the final peptide panel will be locked in a public preregistration. This phase uses work already completed prior to this application, but still requires paid compute for final archival reruns, control generation, reproducibility checks, and assembly of the preregistered candidate package. No candidate choice will be influenced by later assay outcomes.

Candidate prioritization is intended to favor structurally plausible engagement of a biologically relevant myostatin surface rather than treating any detectable surface binding as sufficient.

Peptide synthesis and quality control

Three peptides will be synthesized commercially at small scale. Acceptance criteria are purity greater than 95%, identity confirmed by LC-MS, and sufficient material for repeat binding measurements. If one peptide fails synthesis or QC, a preregistered backup candidate will be substituted only if that substitution rule is specified before any assay results are observed.

Target protein and assay

Binding will be measured against recombinant mature human myostatin using surface plasmon resonance. The assay plan includes matched testing of reference and candidate peptides against the same myostatin preparation, a concentration range appropriate for expected weak-to-moderate peptide binding, repeat measurements for each peptide to confirm reproducibility, and steady-state binding analysis as the primary readout, with kinetic fitting used if supported by data quality.

Direct binding is the critical go/no-go criterion for this pilot. Without clear target engagement, larger functional studies would be conceptually weak even if technically feasible. If a candidate meets preregistered binding criteria, the immediate follow-up gate within this pilot will include a minimal serum or plasma stability assessment at 37°C by LC-MS.

Exact assay parameters, including capture strategy, chip chemistry, concentration window, buffer conditions, replicate depth, and regeneration settings, will be finalized with the contracted SPR facility based on the selected protein format and locked in the public preregistration before data collection begins.

Vendor-generated assay outputs will be reviewed for baseline behavior, reference subtraction quality, replicate agreement, and fit quality against preregistered interpretation rules before any go/no-go conclusion is made.

Specificity check

A limited counter-screen panel will be included to improve interpretability of positive binding results. GDF-11 will be treated as the closest-family cross-reactivity check because of its close relationship to myostatin, while activin A will serve as a broader selectivity comparator within the TGF-β family. This panel is intended to distinguish a narrowly targeted binder from a more broadly cross-reactive peptide, not to establish definitive translational selectivity in a single pilot.

Timeline

Month 1: finalize peptide panel, preregister protocol, place synthesis orders, procure recombinant protein, and schedule assay work.

Months 2 to 3: receive peptides, complete QC, run SPR experiments, and repeat any failed technical runs.

Month 4: run the limited counter-screen panel, complete minimal serum or plasma stability assessment for any positive binder, and perform confirmatory repeat measurements on the top candidate if needed.

Months 5 to 6: complete analysis, release data and methods, and prepare the public report and follow-on plan.

Planned Analyses

This is a pilot study, so the emphasis is on reproducibility and effect size rather than large-N hypothesis testing.

Pre-registered acceptance criteria are:

at least one candidate must show reproducible measurable binding above the negative control across replicate runs
the top candidate should show an apparent Kd no worse than 3-fold above the literature-derived reference peptide under matched experimental conditions
the negative control should show no measurable binding up to the top concentration tested, or binding clearly weaker than the reference/candidate pair
the top candidate should show stronger engagement of myostatin than activin A under matched conditions, while the GDF-11 result will be interpreted as a closest-family cross-reactivity readout rather than a stand-alone selectivity threshold
if none of these conditions are met, the study will be treated as a negative result and reported in full

SPR outputs will be processed using standard reference subtraction and fit pipelines appropriate to the instrument/vendor workflow. Steady-state binding will be the default interpretation route; kinetic fitting will be attempted only if the data quality justifies it. Replicate handling, fitting rules, and outlier treatment will be specified in the public preregistration before synthesis begins.

Results will be interpreted symmetrically. A positive result means at least one computationally prioritized peptide met the preregistered engagement criteria. A negative result means the current design route failed on this target under the tested conditions. Both outcomes are informative.

If the planned peptide panel changes due to synthesis or QC failure, any replacement will come only from a pre-ranked backup list defined before assay execution. Deviations from the original plan will be documented publicly.

Ethics and Data Management

No human tissue, stem cells, animal studies, or in vivo procedures are included in this pilot. Formal human-subjects or animal ethics approval is therefore not required for the proposed work as written.

The study will be preregistered on ResearchHub before experimental work begins. Raw assay outputs, fitted parameters, analysis code, and protocol details will be shared openly regardless of outcome. Data and analysis materials will be archived in openly accessible repositories, and progress updates will be posted during project execution.

4. Pilot Data (Optional)

Internal pilot work has already established the feasibility of this project at the computational level. The target structure, reference scaffold, and candidate-ranking workflow are already in place, and the myostatin program is currently the most mature target in our internal pipeline. This prior work does not substitute for wet-lab validation, but it is the reason this proposal is a focused validation pilot rather than an exploratory target search. If a candidate clears the binding and stability gates in this pilot, the next phase will prioritize orthogonal biophysical confirmation and focused functional cell-based testing.

5. Budget

Item	Cost
Peptide synthesis and QC for 3 peptides	$2,000
Recombinant mature human myostatin	$700
Counter-screen reagents for GDF-11 and activin A	$900
SPR contract service	$4,300
Cloud GPU / model / computational locking costs	$600
Serum or plasma stability assay	$500
Shipping, buffers, and assay consumables	$500
Preregistration, data release, and open repository costs	$200
Contingency	$300
Total	$10,000

Budget estimates are based on current public vendor pricing for peptide synthesis and recombinant proteins, together with anticipated pilot-scale contract service costs for SPR and supporting assays.

If actual vendor pricing is lower than expected, remaining funds will be used to increase replicate depth or strengthen confirmatory control measurements.

6. References

Amthor, H. et al. Lack of myostatin results in excessive muscle growth but impaired force generation. Proc. Natl Acad. Sci. USA 104, 1835–1840 (2007).
Schuelke, M. et al. Myostatin mutation associated with gross muscle hypertrophy in a child. N. Engl. J. Med. 350, 2682–2688 (2004).
Cash, J. N. et al. Structure of myostatin·follistatin-like 3: N-terminal domains of follistatin-type molecules exhibit alternative modes of binding. J. Biol. Chem. 287, 1043–1053 (2012).

A preregistered pilot to validate short peptide myostatin inhibitors for muscle preservation and recovery

A preregistered pilot to validate short peptide myostatin inhibitors for muscle preservation and recovery

Funding Progress

Nonprofit Recipient

Peer Reviews

Topics

DOI