SNSarfaraz K. Niazi

Sarfaraz K. Niazi

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Sarfaraz K. Niazi has a total of 19 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    A Testable Framework for Mechanics-Conditioned Perturbation Models in Drug Discovery

    by Mikołaj Stańczak, Sarfaraz K. Niazi

    Perturbation-predictive cell models are advancing quickly, yet they remain predominantly molecular, even though the mechanical state of cells and tissues—stiffness, viscoelasticity, geometry, and force generation—can alter therapeutic response. This review converts that observation into a testable program. It separates four claims that are routinely conflated and gives each its own baseline, estimand, design, and named refutation condition: whether auxiliary mechanical measurements improve prediction of treatment response beyond a complete non-mechanical baseline (H1); whether mechanics computed from images help as a representation rather than as added information (H2); whether experimentally controlled mechanical context modifies treatment effects, including once local drug exposure is held comparable (H3 and H3a); and whether mechanics-aware predictions improve preclinical prioritization decisions under a fixed budget (H4). Current evidence is graded against an explicit, reproducible rubric, and the two links the framework actually requires, mechanical conditioning of perturbation response and prospectively validated mechanics-aware decisions, are unpopulated. A worked pancreatic ductal adenocarcinoma program specifies donors, materials, gemcitabine exposure measurement, endpoints, simulation-based power, staged costs, and a blinded decision test, with a reduced-scope minimum implementation for groups that cannot execute the full design. The practical contribution for a bioengineering readership is a transferable specification checklist, covering measurement provenance, units of assignment and inference, leakage control, effect margins, and decision rules, to be completed before resources are committed. We conclude that mechanics should enter perturbation models as a hypothesis carrying a refutation condition, not as a default enrichment. Explicit limitations bound the work: the synthesis is targeted rather than systematic, the evidence grades were assigned by a single assessor, and the framework has not itself been executed.

  • Open Access

    From telomere dysfunction to seno-antigen selection: a conceptual framework and route-specific evidence appraisal for senolytic immunotherapy

    by Sarfaraz K. Niazi

    Cellular senescence causally contributes to age-related and chronic diseases, rendering the targeted immune clearance of pathological senescent cells a promising therapeutic avenue. Progress remains limited by heterogeneity across cell lineages, tissues, inducing stressors, and temporal factors. As antigen expression alone does not ensure immune accessibility, functional recognition, or safety for healthy tissues, these aspects warrant thorough evaluation. A conceptual framework is proposed, distinct from a validated or implementable algorithm, for identifying and prioritizing senescence-associated antigens. This framework requires four criteria that a single candidate antigen must meet within the same cellular context: persistent pathological senescence, immune accessibility via a confirmed surface antigen or naturally presented human leukocyte antigen (HLA) peptide, modality-appropriate functional immune recognition, and exclusion from vital healthy or beneficially senescent states. The mechanistic origin, such as telomere dysfunction, serves as an optional stratification axis that defines coherent subsets rather than a universal prerequisite; data lacking matched telomere and antigen information in the same cells remain unverified. The process distinguishes target nomination from product validation: evidence of recognition in a single modality, such as chimeric antigen receptor (CAR) targeting of a surface antigen, does not confirm validation for alternatives like peptide vaccines. Application of the framework to urokinase-type plasminogen activator receptor (uPAR) and glycoprotein nonmetastatic melanoma protein B (GPNMB) indicates that neither candidate currently meets all criteria, and the telomere axis remains untested for both. Given that multivalent vaccination elicits immunity against each component independently, combining candidates that fail individual criteria risks amplifying liabilities rather than ensuring intersectional specificity. This outcome would require an engineered AND-gated recognition system, where the capitalized operators AND, OR, and NOT denote Boolean logic gates rather than emphasis. The framework remains in the validation phase. Each criterion receives an ordinal grade, a non-compensatory decision rule, a route-specific ligand-density window that sets the quantitative condition for recognition and healthy-cell sparing, and a predefined seven-compartment counter-screening panel. Recognition architectures are formally compared, and a pre-registered benchmarking study with specified falsification criteria is outlined as the framework’s validation test.

  • Open Access

    Nontumor‐Specific mRNA Sensitization of Cold Tumors to Immune Checkpoint Blockade: Evidence, Bias, and Requirements for Translation

    by Sarfaraz K. Niazi

    Nontumor‐specific messenger RNA may sensitize cold tumors to checkpoint blockade, but survival estimates are heterogeneous, bias‐aligned analyses reveal no difference, and spreading is unproven.

  • Open Access

    Proportionate evidence in biosimilar development: when a comparative efficacy study can still change the regulatory conclusion

    by Sarfaraz K. Niazi

    Biosimilar regulation was established using a systematic, stepwise framework in which clinical comparisons addressed residual uncertainties following comprehensive quality and pharmacokinetic assessments. Advances in analytical resolution, clinical pharmacology, and over two decades of regulatory expertise have substantially mitigated this residual uncertainty. This publication examines circumstances in which a comparative efficacy study could still influence regulatory decisions. It presents four original analyses: a quantitative comparison of the nominal resolution of selected analytical methodologies, combined with predefined equivalence criteria, within specific clinical programs that operate at disparate scales and are used solely to demonstrate respective resolution and acceptance thresholds. Additionally, an audit of publicly available regulatory outcomes identifies 100 United States 351(k)-licensed biosimilar and interchangeable products, alongside 149 authorized European biosimilars. It also classifies 16 European nonpositive initial outcomes: nine primarily driven by quality, manufacturing, or product bridging issues; two involving clinical pharmacology or overall comparability; one following a positive opinion; and four lacking sufficient public detail. Notably, no case identified a failed comparative efficacy study as the sole decisive factor, a conclusion limited by excluding programs abandoned before filing from the dataset. An analysis of participant burden estimates indicates an aggregate of more than 10,000 enrollments for programs approved through 2019. A Bayesian value of information model shows that, under a defined base case, the net benefit of such a study diminishes once prior evidence indicates that clinical similarity substantially exceeds the decision threshold. The framework continues to emphasize targeted clinical evidence relevant to unresolved mechanisms, uncharacterized local exposure, potential immune risks, or unaddressed population-specific questions.

  • Open Access

    Computational design and translational development of TCR-mimic scFv antibodies for peptide-HLA neoantigen targeting

    by Sarfaraz K. Niazi, Matthias Magoola, Zamara Mariam

    Single-chain variable fragments (scFvs) penetrate tissue more effectively than full antibodies and are simpler to produce, but the mutant proteins that define a tumor are predominantly intracellular and therefore inaccessible to conventional antibodies. What is accessible is the mutation-derived peptide displayed on cell-surface HLA. This review concerns TCR-mimic scFvs directed at those peptide-HLA complexes. It separates three questions that are frequently conflated: which mutations are predicted to be presented, which presented complexes are genuinely targetable given surface density and cross-reactivity risk, and how immunogenic the resulting therapeutic molecule will itself be. We review computational strategies spanning target identification, AI-driven structure prediction, de novo design, affinity and specificity engineering, half-life extension and deimmunization, and distinguish throughout between the scFv as a recognition module and the effector architecture, such as a CD3 bispecific, a chimeric antigen receptor binding domain or an Fc-containing format, in which it must ultimately be deployed.

  • Open Access

    Drugging the Undruggable in Oncology: Structural Computation, Machine Learning, and Generative AI for Historically Intractable Targets

    by Mikołaj Stańczak, Sarfaraz Niazi

    For four decades the central oncogenic drivers of human cancer the RAS GTPases, the transcription-factor MYC, mutant p53, and the protein-tyrosine phosphatases, fusion oncoproteins and intrinsically disordered regulators that surround them were regarded as undruggable: lacking deep, well-defined small-molecule binding pockets, governed by femtomolar substrate affinities, or driven through flat protein–protein interfaces that classical medicinal chemistry could not engage. The last decade has overturned that verdict. This narrative review argues that the convergence of three advances has made the undruggable proteome systematically approachable: the discovery of cryptic, allosteric and induced pockets through molecular dynamics and deep learning; the maturation of physics-based and machine-learned models that can read out binding from structure; and the rise in generative artificial intelligence that can design novel molecules conditioned directly on a three-dimensional target. We develop the argument in full formal detail. We set out the mathematics of modern molecular machine learning E(3)/SE(3)-equivariant graph neural networks, denoising diffusion and score-based generative models, and generative flow networks alongside the statistical mechanics of binding free energy (free-energy perturbation, thermodynamic integration, MM-PBSA), the kinetics of covalent inhibition, the equilibrium theory of ternary complexes and cooperativity that governs molecular glues and degraders, and the pharmacology that determines whether a designed molecule reaches and acts on a tumour. We then examine, target by target, the biology that made each oncoprotein intractable and the campaigns structural, chemical and increasingly computational that have begun to defeat it, distinguishing throughout what classical structural and computational work achieved, what machine learning demonstrably contributed, and what generative design has so far only promised: the switch-II pocket and the covalent KRAS-G12C and non-covalent KRAS-G12D and pan-RAS(ON) inhibitors; the mutation-created cavity of p53-Y220C and its first reactivators; the disordered MYC-MAX interface and the miniprotein Omomyc; and the allosteric tunnel of SHP2. These four are selected for mechanistic non-redundancy, on grounds we state explicitly, and the structured search framework and eligibility criteria underlying the review are documented. We foreground the newest evidence AlphaFold3, RoseTTAFold All-Atom, deep cryptic-pocket predictors, and the first generative-to-clinic existence proofs while maintaining candour about the field’s hard limits: the physical-validity and generalisation gaps of deep docking, the bounded accuracy and cost of free-energy methods, the synthesizability gap, dual-use risk, and a clinical attrition that artificial intelligence has not yet reduced. The synthesis we offer is not that the problem is solved but that it has changed in kind: from a search for the rare tractable target to the deliberate, computable engineering of molecules against targets once thought beyond reach.

  • Egalitarianism as hierarchy control: Predictive social minds and the evolution of anti-dominance institutions

    by Sarfaraz K. Niazi

    Stibbard-Hawkes and von Rueden redefine egalitarianism as a process rather than an outcome. We extend this framework by arguing that egalitarian societies represent cognitive-ecological equilibria in which evolved psychological systems regulate dominance hierarchies. The six identified processes correspond to distinct neurocognitive mechanisms for hierarchy suppression, interpretable as distributed strategies for managing hierarchical risk within predictive processing.

  • Open Access

    Reference-Anchored Equivalence: A Unified Scientific Framework for the Approval of Follow-On Drugs and Biological Products

    by Sarfaraz K. Niazi

    Follow-on approval processes in the United States are conducted through three distinct legal pathways: abbreviated new drug applications under section 505(j) of the Federal Food, Drug, and Cosmetic Act; applications under section 505(b)(2); and biosimilar applications under section 351(k) of the Public Health Service Act. A determination made under one pathway does not serve as a precedent for another. Nevertheless, these regulatory regimes rest on a shared scientific principle known as reference-anchored equivalence: reliance on a reference product is justified when sufficiently sensitive comparative evidence shows no clinically meaningful differences within a legally permissible comparison, while the successor product’s inherent risks are independently assessed. The FDA’s August 2026 therapeutic equivalence guidance illustrates the agency’s current application of product-level, evidence-based reasoning, particularly when identity cannot be presumed. A comprehensive framework is proposed, including a legal eligibility gate, a reference sampling gate, six evidence domains, pre-established decision criteria, independent assessment of successor-specific risks, and explicit cessation protocols. This framework has been tested against a hypothetical scenario involving a same-serotype, same-cassette adeno-associated virus successor, integrated within an operational pathway based on existing procedural mechanisms. Distinguish scientific unification from legal consolidation: products whose capsid, lipid nanoparticle, or cellular component directly contribute to clinical function may fall outside the statutory definition of biosimilarity and, consequently, require novel reliance authority.

  • Open Access

    Explicit Mechanistic Operators in Predictive Virtual Cells: Evidence, Failure Modes, and a Minimum Falsification Framework for Single-Cell Perturbation Prediction

    by Mikołaj Stańczak, Sarfaraz K. Niazi

    The term virtual cell now covers objects ranging from curated biochemical simulators to atlas-trained neural representations, yet the decisive empirical question is narrower: whether a model built to obey known biochemical rules predicts the effects of new drugs or genetic changes better than an otherwise identical model that lacks them. Formally, this Perspective asks whether a biologically specified mechanistic operator improves out-of-distribution prediction of single-cell perturbation responses beyond strong simple, relational, transport-based, and learned-dynamical comparators. Recent benchmarks show that rankings depend on partition, statistical unit, effect-size distribution, gene set, and metric, and that several deep and foundation-model approaches fail to exceed no-change, mean-effect, additive, or linear controls. Set-based, knowledge-graph, transport, and learned state-equation models can improve defined tasks. No identified study isolates the incremental value of a biologically specified operator against late fusion, a capacity-matched static model, a capacity-matched learned-dynamical operator, and topology-rewired controls under one locked evaluation. Six source categories are distinguished, with the operator category divided into biologically specified and learned-dynamical subclasses. A minimum falsification framework is specified for the unfolded protein response with a declared capacity ledger, endpoint times chosen by a prespecified Stage 1 rule, and a locked comparator set. The framework is prospective: no new data, model implementation, or performance result is reported, and every threshold is a design choice awaiting empirical test.