How to Develop CDK4-Selective PROTACs: Warhead Selection, Linker Design, and E3 Ligase Selection

* Please be kindly noted that our services and products are intended for research use by organizations or companies and are not intended for individual use.

Consult with Our Experts

What Is CDK4, and Which CDK4 Forms Are Worth Targeting with Degraders?

Cyclin-dependent kinase 4 (CDK4) is a serine/threonine kinase that partners with cyclin D1, D2, and D3 to drive the G1-to-S transition. Once the cyclin D–CDK4 complex is assembled and Thr172 in the activation loop is phosphorylated, the complex phosphorylates Rb family proteins, releasing E2F transcription factors and committing the cell to replication. Because this node sits directly upstream of cell-cycle entry, CDK4 is one of the most intensively studied kinases in the field.

Structure matters more here than for many kinase targets. CDK4 contains roughly 300 amino acid residues and adopts the classic bilobal kinase fold: an N-lobe carrying the αC helix, a C-terminal helical lobe, a hinge, a glycine-rich loop, and an activation loop containing Thr172. Monomeric CDK4 is not catalytically competent. Cyclin D binding remodels the αC helix and activation loop, and phosphorylation then completes activation. A second input comes from p16 (CDKN2A), which binds the face opposite the cyclin D interface and distorts the ATP pocket rather than blocking it. The pocket geometry therefore depends on whether CDK4 is monomeric, cyclin-bound, phosphorylated, or held in a p16 complex.

For degrader work, that heterogeneity is an advantage. An inhibitor has to occupy the pocket in a way that blocks catalysis, so its activity depends on the conformational state. A degrader only has to bind: any site supporting a productive ternary complex can support removal, including states in which the pocket is distorted. Removal also clears the non-catalytic CDK4 pool that sequesters p21 and p27 and thereby influences CDK2 activity, which inhibition does not address.

The demand for CDK4 selectivity is structural. The catalytic domains of CDK4 and CDK6 are closely related and their ATP pockets are nearly superimposable, which is why inhibitors rarely separate the two. CDK6 is highly expressed in hematopoietic lineages while CDK4 is more broadly distributed, so the proteins are not interchangeable in every model. A degrader that removes one paralog and spares the other lets a team attribute biology without relying on selectivity the ATP site does not offer.

Table 1. CDK4 Forms and Their Implications for PROTAC Design.

Developed CDK4-Selective PROTACs and Their Research Progress

Nearly every reported CDK4 degrader starts from an established CDK4/6 inhibitor chemotype. That is a sensible shortcut: the binding mode is published and co-crystal structures show where a linker can leave the pocket without breaking key contacts. The published work also shows where the shortcut stops paying off. The warhead sets the floor for engagement, while linker architecture and recruiter choice decide which paralog is removed and how clean the profile is.

Table 2. Reported CDK4 and CDK4/6 Degraders and Their Research Findings.

Four transferable lessons follow, and each maps onto a design decision in the sections below.

• Selectivity is built, not inherited: these warheads bind CDK4 and CDK6 with comparable affinity, so paralog preference has to be engineered into the ternary complex.
• CDK6 is the easier default: several series degraded CDK6 more strongly than CDK4, attributed to a more stable ternary complex or faster CDK4 recovery. A CDK4 profile requires deliberate design.
• Short and rigid is the recurring CDK4 pattern: the clearest CDK4-selective examples used a short alkyl linker on a ribociclib-type warhead, while an extended PEG linker on a palbociclib-type warhead inverted the preference.
• Liabilities are inherited: an imide recruiter adds IKZF1 and IKZF3, and an abemaciclib-type warhead brings CDK9. Both are fixed at their source, not downstream.

ATP-Site-Directed CDK4 PROTAC Development with Palbociclib-Type Warheads

Target Site: CDK4 ATP Pocket and the Solvent-Exposed Piperazine Exit Vector

Co-crystal structures of palbociclib, ribociclib, and abemaciclib with CDK6 (PDB entries 5L2I, 5L2T, and 5L2S) explain why this chemotype is productive. The aminopyrimidine forms hinge hydrogen bonds, and the piperazine ring is held against a solvent-exposed ridge formed by an aspartate and a threonine. CDK4 presents the equivalent architecture. The key structural fact is directional: the piperazine projects out of the cleft toward solvent while the rest of the inhibitor stays buried.

That geometry defines the exit vector. Derivatizing the piperazine 4-nitrogen extends into solvent instead of into the pocket wall, which is why potency survived derivatization across the published series. Choose the attachment point from structure and then verify binding, rather than selecting a convenient handle and hoping the binding mode survives.

Warhead Selection: Aminopyrimidine Inhibitor Scaffolds and Affinity Requirements

Palbociclib-type scaffolds give published binding modes, well-characterized hinge interactions, and an obvious solvent-exposed handle, and conjugates retained low-nanomolar IC₅₀ values after derivatization. Two cautions apply. Retained enzymatic potency does not predict cellular degradation, because ternary complex formation and exposure sit downstream. And these warheads engage CDK4 and CDK6 similarly, so a palbociclib-type warhead alone will never deliver paralog selectivity.

• Keep the hinge contacts intact: the aminopyrimidine hydrogen bonds and the piperazine–ridge contacts anchor the binding mode, so derivatization should extend from them rather than replace them.
• Measure the intermediate: compare the parent, the linker-bearing intermediate, and the final degrader. Loss at the intermediate stage indicates an exit-vector problem.
• Set an affinity floor, not a goal: low-nanomolar potency is sufficient. Extra affinity costs molecular weight and polarity without improving D_max.

Selecting and validating the attachment position is a separate step, handled by our linker binding site selection and design work before any linker library is built.

Linker Design: Alkyl versus PEG Chains, Chain Length, and Attachment Chemistry

The literature is unusually clear that length and composition change which paralog is removed. With a palbociclib-type warhead and a CRBN recruiter, an alkyl-linked compound degraded both CDK4 and CDK6, while a PEG-3 extension shifted the same warhead toward selective CDK6 removal. A separate series reached selective CDK6 loss with the longest linker and only partial loss with shorter ones.

Attachment chemistry deserves separate treatment. In one systematic study, the linker junction on the warhead side — amide, triazole, or methylene — had little effect on activity, whereas the recruiter-side junction did matter. Conjugation chemistry can therefore follow synthetic convenience on the warhead side while the recruiter-side junction stays a design variable.

• Scan length as a matched series: hold the warhead and recruiter constant and vary only the chain, from short alkyl through PEG-2 to PEG-4.
• Choose composition for the missing property: alkyl chains favor permeability and reduce polar surface area; PEG chains improve solubility and reach. With an already polar warhead, an alkyl-first sweep is usually the better start.
• Read the hook effect as information: when degradation falls at high concentration, the linker favors binary over ternary complexes. That is a geometry problem, not a potency problem.

E3 Ligase Selection: CRBN Recruitment and Control of IKZF1/IKZF3 Co-Degradation

CRBN is the default recruiter because imide chemistry is well developed, ligand-linker conjugates exist in many lengths, and the best-characterized CDK4/6 degraders recruit it. The cost is neosubstrate activity: IKZF1 and IKZF3 are degraded independently of CDK4 engagement, and several reported CDK4 degraders reduced them as well, which changes the phenotype in hematological models.

• Include a modified-imide comparator: replacing the aryl amine with an oxylamide reduced IKZF1/3 recruitment in a published CDK4/6 series, which is the cleanest way to separate CDK4-driven from neosubstrate-driven effects.
• Confirm recruiter dependence: competition with the free recruiter, or loss of degradation in a recruiter-deficient model, establishes that removal runs through the intended ligase.
• Re-measure properties after changing recruiter: moving between pomalidomide, lenalidomide, and shorter imides changes polarity and stability of the whole molecule.

Table 3. Palbociclib-Type CDK4 Degrader Design Summary.

CDK4-Selective PROTAC Development with Ribociclib- and Abemaciclib-Type Warheads

Target Site: Subtle ATP-Site and Surface Differences Between CDK4 and CDK6

CDK4 and CDK6 cannot be separated inside the ATP pocket, because the residues lining the cleft are effectively conserved. Selectivity has to come from the ternary complex interface, where contacts involve the N-lobe, the αC helix region, and loops that differ more between the two paralogs than the pocket itself. Small differences on those peripheral surfaces are amplified into measurable differences in complex stability.

There is also a dynamic component. CDK4 and CDK6 differ in how readily they form a productive complex and how fast the target pool recovers. The two published explanations for CDK4-sparing profiles — a less stable CDK4 ternary complex, or faster reversal of CDK4 ubiquitination — predict the same outcome, so measure assembly and recovery directly rather than inferring mechanism from a blot.

Warhead Selection: Ribociclib-Type Scaffolds and the CDK9 Liability of Abemaciclib-Type Scaffolds

The clearest CDK4-selective degraders come from ribociclib-type warheads. A ribociclib-based compound with a four-carbon alkyl linker degraded CDK4 selectively, and a later optimized compound removed CDK4 without measurably reducing CDK6 or IKZF1/3, with cereblon-dependent degradation. In another series, ribociclib-containing PROTACs did not degrade CDK6 at all.

Abemaciclib-type warheads deserve a specific warning. Abemaciclib has broader kinase activity, and conjugates built on it degraded CDK9 alongside CDK4 and CDK6. CDK9 degradation changes transcriptional regulation, so it is not a benign addition to the profile.

• Default to a ribociclib-type warhead for CDK4 selectivity: the published CDK4-selective examples cluster there and the effect has now been reproduced in more than one series.
• Treat CDK9 as a warhead-inherited risk: include CDK9 in the selectivity panel whenever an abemaciclib-type chemotype is used, and report it as a degradation off-target.
• Do not over-read enzymatic potency: the warhead sets engagement, while paralog preference is decided later, so a weaker binder can still become the better CDK4-selective degrader.

Linker Design: Short Alkyl Linkers and Conformational Restriction for CDK4 Preference

The most useful comparison in this area pairs a ribociclib-type warhead with a four-carbon alkyl linker, which favored CDK4, against a palbociclib-type warhead with an extended PEG-3 linker, which favored CDK6. A short, hydrophobic chain appears to favor the more compact CDK4-presenting geometry, while a longer, flexible chain gives CDK6 enough freedom to form a stable complex. Because warhead and linker were changed together in these examples, treat the pattern as a hypothesis for a designed matrix rather than a rule.

• Keep the first CDK4 matrix short: two- to six-carbon alkyl chains are the most informative first sweep when CDK4 selectivity is the goal.
• Pair every alkyl analogue with an equal-length PEG comparator: this separates a length effect from a polarity or flexibility effect, which is otherwise impossible to attribute.
• Add rigidity as a second axis: once an active length is found, insert a piperazine, piperidine, pyridine, or alkyne unit, because a rigid analogue sometimes rescues a length that failed as a flexible chain.

E3 Ligase Selection: Imide Variants, VHL, and IAP Recruiters in Selectivity Profiling

Recruiter choice is the third lever, and the evidence supports treating it as a variable. In one study, palbociclib-type warheads were connected through the same piperazine attachment to CRBN, VHL, and IAP recruiters. All three produced CDK4 and CDK6 degradation, with reported pDC₅₀ values of roughly 6.2–8.0 for CDK4 and 7.7–9.1 for CDK6, and the VHL- and IAP-based compounds were the least potent. Ligase identity changes how much degradation is achieved but does not by itself create CDK4 selectivity.

The comparison is still worth running for two other reasons. VHL ligands are larger and more polar, which shifts solubility and permeability in a predictable direction, and they carry no IKZF1/3 liability. Where the primary problem is an imide-driven off-target, switching to VHL is often the faster fix.

• Run two recruiters in the first matrix: comparing CRBN and VHL early costs less than rebuilding a linker series around the wrong recruiter later.
• Use VHL when the imide is the problem: a VHL ligand-linker conjugate removes the neosubstrate liability without touching the warhead or linker.
• Keep IAP and MDM2 as comparators: both have published CDK4/6 precedent, but each brings its own biology that has to be controlled for.

Table 4. CDK4-Selective Degrader Design Summary for Ribociclib- and Abemaciclib-Type Warheads.

Cyclin D1-CDK4 Complex-Targeted PROTAC Development

Target Site: Cyclin D-CDK4 Interface and Complex-Specific Surfaces

Cyclin D1 is a compelling partner target and a difficult one. It drives CDK4 activation and its gene is amplified in several tumor types, but it has no conventional small-molecule pocket. The cyclin D–CDK4 interface is large and relatively flat, the classic profile of a protein–protein interaction surface.

The complex creates something neither protein offers alone: adjacent surfaces that exist only when the two are assembled. If a small molecule binds one subunit and recruits an E3 ligase, the ubiquitination machinery is brought near the whole assembly, and which subunit is modified depends on geometry rather than on which subunit carries the ligand. That is the basis of the bridged PROTAC concept.

Warhead Selection: Complex-Binding Ligands and Partner-Protein Recruiters

The instructive example is a bridged degrader built from a CDK4/6-binding warhead and a VHL-recruiting ligand. Instead of targeting cyclin D1 directly, it binds CDK4/6 and pulls the assembly toward VHL. The reported result was degradation of cyclin D1 ahead of CDK4 and CDK6, greater potency than CDK4/6 inhibition or CDK4/6 degradation, and activity in triple-negative breast cancer lines that were largely insensitive to either.

• Bind the tractable subunit: a CDK4/6 ATP-site ligand is sufficient as the complex anchor, and no cyclin D1 binder is required.
• Decide which subunit should fall first: degradation order is a design outcome, so state the intent before screening and measure the right protein.
• Expect complex-specific phenotypes: cyclin D1 removal produced a transcriptional signature not seen with CDK4/6 inhibitor or degrader treatment, so do not copy inhibitor protocols for downstream readouts.

Linker Design: Orientation, Spacing, and Geometry for Complex-Level Degradation

Geometry carries more weight here than in a single-subunit degrader. A 2025 DNA-templated study made this explicit: controlling the distance and orientation between the CDK4/6-binding element and the recruiter produced constructs with distance- and orientation-dependent subunit selectivity, and the optimized construct achieved synchronous degradation of the whole cyclin D1–CDK4/6 complex. Spacing and orientation are separable variables, and changing either one changes which subunit is ubiquitinated.

• Vary one axis at a time: change distance at fixed geometry, then geometry at the active distance. Changing both together makes the result uninterpretable.
• Use rigid elements to define geometry: aromatic, heteroaryl, alkyne, and piperazine units constrain the exit angle, which determines which lysines on the complex are reachable.
• Measure every subunit: quantify cyclin D1, CDK4, and CDK6 in the same experiment, because selective loss of one subunit is easy to miss if only one protein is probed.

E3 Ligase Selection: VHL Recruitment and Alternatives for Complex Degradation

VHL has the strongest precedent for complex-level work: the bridged cyclin D1 degrader and the templated complex series both recruit it. Whether that reflects a genuine geometric advantage or the choices made in those programs is unsettled, so keep the CRBN comparison in the design.

• Start from VHL for complex-level designs: both published complex-directed CDK4 programs used it, and the conjugates are available in several lengths and handle types.
• Confirm the mechanism: competition with the free recruiter and a non-degrading control establish that removal of each subunit runs through the intended ligase.

Table 5. Cyclin D1-CDK4 Complex-Targeted Degrader Design Summary.

CDK4 PROTAC Development for Altered-Target and Inhibitor-Resistance Contexts

Target Site: Amplified CDK4, p16-Insensitive CDK4 Mutants, and Altered Complex Pools

Resistance to CDK4/6 inhibition converges on the same node and mostly increases CDK4 signaling rather than changing CDK4. RB1 loss removes the requirement for CDK4 activity, CCNE1 amplification activates CDK2 as a parallel route, and CDK6 or CDK4 amplification raises the amount of kinase that must be suppressed. CDK4 amplification in the 12q13–15 region, frequently co-amplified with MDM2, raises the target burden directly.

Two CDK4-specific alterations deserve separation. The R24H and R24C substitutions make CDK4 insensitive to p16 while leaving the ATP site intact, so an inhibitor has less to work with while a degrader is unaffected. The same logic applies to CDKN2A loss, which raises the active CDK4 fraction without altering the pocket.

Warhead Selection: Ligands That Stay Engaged When ATP-Site Occupancy Is Not the Limiting Factor

When the target pool is large, occupancy stops being the bottleneck. Failures in this setting are throughput problems: insufficient ternary complex, insufficient ubiquitination, or resynthesis faster than removal. That reframes warhead selection toward binders that form a productive and catalytically efficient ternary complex, because the degrader is reused rather than consumed.

• Rank on turnover, not affinity: compare candidates on D_max and on the rate of degradation at fixed concentration, not on IC₅₀ alone.
• Keep the ATP-site warhead: the pocket is intact in the great majority of resistance mechanisms, so inhibitor-derived chemotypes remain the right starting point.
• Confirm engagement on the altered form: measure target levels before interpreting degradation depth, since amplification changes how much CDK4 is present.

Linker Design: Sustaining Degradation Efficiency at High Target Abundance

At high abundance the linker has to support turnover. Ternary complex stability determines how many degradation events each molecule delivers, while cell exposure determines whether enough compound reaches a larger pool. These two requirements pull in opposite directions, because the rigid and polar elements that stabilize the assembly usually reduce permeability — which is why the linker must be optimized on the complete molecule.

• Optimize for residence in the ternary complex: compare analogues on the stability of the assembled complex, not only on degradation depth at one time point.
• Protect exposure first: in an amplified model, a compound that reaches the target at lower concentration can still give a deeper effect.
• Keep the molecule as small as the geometry allows: every additional linker atom is paid for in permeability, so use the shortest productive length.

E3 Ligase Selection: Matching E3 Recruiters to Cell Models with Altered CDK4 Turnover

Ligase choice interacts with the model in two ways. Recruiter abundance differs between cell systems, and a resistant derivative line may not match its parent. Target turnover also matters, because CDK4 recovery depends on resynthesis and on reversal of ubiquitination, and different recruiters place ubiquitin on different surfaces with different consequences for duration.

• Change one variable: swap the recruiter while holding warhead and linker constant to isolate the ligase contribution.
• Read durability, not only depth: measure CDK4 at several times after compound removal to separate slow degradation from fast resynthesis.
• Keep the paralog readout: CDK6 levels in the same sample show whether the selectivity window survived the change of model.

Table 6. CDK4 Degrader Design Summary for Altered-Target and Resistance Contexts.

Common Failure Causes and Optimization Strategies for CDK4-Selective PROTACs

Weak or Non-Productive Ternary Complex Formation

The most common CDK4 degrader failure is not poor binding. It is a compound that binds CDK4 and its ligase well in isolation but never assembles the three components productively. Because CDK4 and CDK6 differ least where the warhead binds, the peripheral contacts that stabilize the complex carry most of the burden, and they are the part most likely to be wrong. The published CDK4/6 work supports this: in several series the CDK6 complex was the more stable one and CDK4 degradation the weaker outcome, without any corresponding affinity difference.

• Measure the complex before making more compounds: our PROTAC ternary complex assay distinguishes an assembly failure from a downstream processing failure, and those have different fixes.
• Change the recruiter first where assembly is weak: moving between CRBN and VHL changes the interface more than a two-atom linker change does.
• Re-examine the exit vector: if assembly fails across a whole length series, the attachment point is probably presenting the recruiter to the wrong face.

Hook Effect and Incomplete, Non-Durable CDK4 Depletion

The second cluster appears only in dose–response data. Above an optimal concentration, degradation falls because binary complexes outcompete the ternary complex. CDK4 programs are prone to this because the warheads are potent and saturate early. The related problem is shallow, short-lived depletion: CDK4 is actively resynthesized, and the published explanation for several CDK4-sparing profiles was faster reversal rather than weaker binding.

• Always run a full concentration series: a single high concentration will find the hook effect and report it as inactivity.
• Report D_max alongside DC₅₀: a potent but shallow degrader is often less useful than a moderately potent one with high D_max. Depth and kinetics are measured together in our degradation ability assay.
• Quantify both paralogs across the whole curve: selectivity can be concentration-dependent, and a window open at one concentration may close at another.

Poor Permeability, Solubility, and Metabolic Stability of Large Degraders

CDK4 degraders sit at the difficult end of property space. The parent inhibitors are already relatively large and polar, and adding a linker and recruiter pushes the molecule well beyond conventional small-molecule space. The familiar failure is a compound with excellent biochemical and ternary behavior that never reaches intracellular CDK4 at a useful concentration.

• Diagnose before optimizing: measure permeability and solubility separately, because a compound can fail on either and the fixes are opposite.
• Reduce exposed hydrogen-bond donors: donor count usually limits permeability in degrader series, and it can often be lowered by N-methylation or by changing a junction chemistry.
• Check stability early: imide recruiters and some linker junctions are metabolically labile. Aqueous handling is tested through solubility and stability testing and liability through PROTAC in vitro metabolism profiling.

Off-Target Degradation of CDK6, CDK9, and IKZF1/IKZF3

Three off-targets account for most CDK4 selectivity problems, each with a different cause and fix. CDK6 is the structural near-neighbor, so the question is whether the ternary complex forms on both. CDK9 is a warhead-inherited liability, prominent for abemaciclib-type chemotypes, and it changes transcription rather than the cell cycle. IKZF1 and IKZF3 are recruiter-inherited neosubstrates of imide CRBN ligands, independent of the target entirely.

• Define the CDK6 window explicitly: report CDK4 and CDK6 DC₅₀ values from the same experiment, and do not describe a dual degrader as CDK4-selective.
• Fix CDK9 at the warhead and IKZF1/3 at the recruiter: switching away from an abemaciclib-type chemotype is more reliable than engineering the off-target away downstream, and a modified non-imide recruiter removes the neosubstrate without touching the warhead.
• Use the right profiling depth: select paralog and neosubstrate questions are resolved by PROTAC selectivity evaluation, and broader questions by PROTAC off-target evaluation.

BOC Sciences Support for CDK4-Selective PROTAC Development

CDK4 programs fail at predictable points: the exit vector is chosen for synthetic convenience, the linker series is scanned at a single recruiter, CDK6 is measured later than it should be, and CDK9 or IKZF1/3 is found after the compound has been optimized. BOC Sciences supports each of those decision points, from CDK4 form assessment through warhead and linker design, synthesis, and paralog-resolved evaluation.

Table 7. Representative Reagents and Building Blocks for CDK4 Degrader Programs.

CDK4 Target Selection and Mutant-Form Degradability Assessment

• Define which CDK4 forms are present: monomeric, cyclin D-bound, p16-bound, p16-insensitive mutant, or amplified
• Map candidate binding surfaces by form and confirm which conformations support an exit vector toward bulk solvent
• Assess the CDK6 context in the same model, since the selectivity window depends on what is measurable there
• Characterize form and abundance through target protein services, with conformational behavior examined by molecular dynamics simulation
• Produce CDK4, cyclin D1, and complex material through custom protein expression and purification

CDK4 Warhead Selection and Target-Ligand Design

• Triage palbociclib-, ribociclib-, and abemaciclib-type chemotypes against the defined CDK4 form and intended paralog profile
• Identify solvent-exposed exit vectors from CDK4 and CDK6 structural information and design multiple attachment-site analogues
• Flag warhead-inherited liabilities such as CDK9 before a library is built, using virtual screening and molecular docking for protein-ligand
• Design linker-functionalized CDK4 ligand intermediates through ligand design for target protein, including small molecule target protein ligand optimization

Linker and E3 Ligase System Design with Custom PROTAC Synthesis

• Build focused matrices varying attachment site, length, composition, and rigidity, drawing on alkyl, PEG, aromatic, heterocyclic, and click-compatible linkers from the linker library
• Generate length series rapidly by click assembly, supported by click chemistry linker design services
• Supply recruiter chemistry for both ligase systems through CRBN ligand design services and VHL ligand design services
• Prepare parent warhead, non-degrading recruiter control, and matched-linker controls, then support progress into PROTAC manufacturing services

CDK4 Degradation, Selectivity, and Mechanistic Evaluation

• Confirm binary binding after derivatization with binding affinity measurement
• Verify target ubiquitination through protein ubiquitination services and the TR-FRET in vitro ubiquitylation assay
• Quantify CDK4 and CDK6 degradation separately with full concentration ranges, D_max, and washout recovery
• Extend promising candidates into exposure work through PROTAC in vivo PK evaluation in animal models

Explore featured products that can expand your research options and accelerate your next discovery.

Access end-to-end service solutions that help bring efficiency, flexibility, and expertise to your research pipeline.

News

• Degrader Building Blocks for Targeted Protein Degradation
• Recent Progress of PROTACs Targeting Non-enzymatic Functional Proteins
• Discovery of Molecular Glue Degraders
• The Next Frontier of PROTACs: Development of Molecular Glues

Technical Information

• Proteolysis-Targeting Chimeras (PROTAC)
• Overview of 12 Targeted Protein Degradation Technologies
• PROTACs vs. Traditional Small Molecule Inhibitors: A Comparative Analysis
• Common Binding Sites for Protein Degrader Design in Kinases: Types, Features, and Validation