DAMPER-Essentiality Decoupling: Why the Network's Most Structurally Disruptive Node Is Not the Most Essential One

The Prediction and Why It Was Pre-Registered

D21 established that SOS1 is the top structural DAMPER in KRAS-mutant LUAD (z_dr=+4.184). SOS1 is the guanine nucleotide exchange factor that catalyzes GDP-to-GTP exchange on RAS proteins -- the upstream activator of KRAS signaling.

The structural prediction was explicit: SOS1 is maximally disruptive in KRAS-MT networks. Remove SOS1 from the network and the co-expression structure changes more than for any other gene. This is a strong signal. The question D22 asked: does this structural signal correspond to genetic essentiality? Are KRAS-mutant cancer cells more dependent on SOS1 for survival?

H1 was pre-registered before loading any DepMap CRISPR data: SOS1 CERES score lower (more essential) in KRAS-MT vs KRAS-WT lung cancer cell lines.

Pre-reg hash: 07da797 (D22_DEPMAP_CRISPR_SOS1_KRAS_v1).

The Result: Opposite Direction

CRISPR essentiality across 22 KRAS-MT and 21 KRAS-WT LUAD cell lines from DepMap:

| Gene | mean CERES, KRAS-MT | mean CERES, KRAS-WT | Cohen d | p | |------|--------------------|--------------------|---------|---| | SOS1 | -0.192 | -0.328 | +0.52 | 0.29 | | RAF1 | -0.149 | -0.074 | -0.29 | 0.84 | | SHP2 | -0.609 | -0.801 | +0.44 | 0.27 | | KRAS | -1.359 | -0.568 | -1.34 | 0.0002 |

SOS1 in KRAS-WT cells: mean CERES = -0.328. More essential. SOS1 in KRAS-MT cells: mean CERES = -0.192. Less essential.

H1 DENIED. The direction is reversed from the pre-registered prediction.

Positive control H4: KRAS is strongly essential in KRAS-MT (d=-1.34, p=0.0002). The annotation is correct; the dataset is clean.

Why This Makes Biological Sense

SOS1 is a RAS guanine nucleotide exchange factor. Its job: catalyze the GDP->GTP exchange that activates RAS. In KRAS-WT cells, this catalytic function is required -- without SOS1, RAS stays GDP-bound (inactive), and the entire MAPK/PI3K downstream axis is quiescent. Knocking out SOS1 in KRAS-WT cells blocks RAS activation. Cells stop growing.

In KRAS-G12X cells, the mutation locks KRAS in the GTP-bound state regardless of SOS1 activity. The G12X substitution impairs intrinsic GTP hydrolysis and prevents GAP-mediated hydrolysis. KRAS stays active constitutively. Knocking out SOS1 in KRAS-MT cells has reduced effect because the oncogenic driver is already constitutively active and no longer depends on GEF input.

Result: KRAS-WT cells are more CRISPR-dependent on SOS1 than KRAS-MT cells. This is not paradoxical -- it is the expected consequence of the mutation's mechanism.

SHP2/PTPN11 shows the same pattern (d=+0.44, p=0.27): also an upstream RAS activator, also less essential in KRAS-MT. The pattern is consistent across multiple upstream RAS activators -- constitutive KRAS activation creates a bypass for all upstream components simultaneously.

Then Why Does BI-3406 Work Selectively in KRAS-MT Cells?

BI-3406 is a clinical-stage SOS1 inhibitor (Boehringer Ingelheim, Phase 1) that shows selectivity for KRAS-mutant cancers -- which appears to contradict our CRISPR result.

The resolution: pharmacological inhibition and genetic knockout are mechanistically different perturbations.

CRISPR knockout removes all SOS1 protein. This eliminates GEF catalysis, scaffolding functions, protein-protein interaction surfaces, and allosteric regulation simultaneously.

BI-3406 occupies the SOS1 catalytic groove without removing the protein. The occupied SOS1 can still bind KRAS-GTP (at the allosteric second site on SOS1). In KRAS-MT cells, KRAS-G12X is constitutively GTP-loaded -- it floods SOS1's allosteric site and creates a prolonged SOS1/KRAS-GTP complex. The KRAS-G12X mutation impairs GTP hydrolysis and prevents efficient complex dissociation. This traps the SOS1 protein in a non-productive state that sequesters KRAS-GTP, exacerbating the inhibitory effect.

In KRAS-WT cells: the SOS1/RAS complex dissociates normally after catalysis. BI-3406 blocks GEF activity but does not create the extended complex.

The KRAS-MT selectivity of BI-3406 is a pharmacological gain-of-function from the mutation, not a reflection of greater SOS1 genetic essentiality. CRISPR measures the null -- what happens without the gene. BI-3406 measures the occupied state -- what happens when the protein is present but catalytically blocked and allosterically engaged.

The DAMPER-Essentiality Decoupling Principle

D22 establishes a general principle for the framework:

Structural DAMPER status predicts pharmacological vulnerability via allosteric disruption. Structural ANCHOR status predicts genetic essentiality. These are orthogonal axes.

| Structural class | Genetic essentiality (CRISPR) | Pharmacological sensitivity | |-----------------|------------------------------|-----------------------------| | Strong ANCHOR | High (driver addiction) | High (TKI/targeted therapy) | | Structural DAMPER | Variable -- may be LOW | High (allosteric inhibitors) | | Near-zero | Background | Background |

SOS1: DAMPER (z=+4.18) -- pharmacological target (BI-3406 KRAS-selective). NOT CRISPR-essential in KRAS-MT. KRAS: partial ANCHOR (z=-0.18) -- CRISPR essential (d=-1.34). Confirmed. EGFR: strong ANCHOR (z=-2.05) -- TKI essential (clinical standard of care). Confirmed by D20.

The structural classification correctly identifies the axis of vulnerability. The type of drug that should be used to exploit a DAMPER is different from the type of drug for an ANCHOR.

The RAF1 Bypass in BRAF-V600E Melanoma

An unexpected finding emerged in the cross-cancer control (H3 test):

RAF1 (CRAF) is significantly less essential in BRAF-V600E melanoma cell lines (n=17) vs BRAF-WT (n=9):
• mean CERES BRAF-V600E: +0.009 (near non-essential)
• mean CERES BRAF-WT: -0.517 (essential)
• Cohen d = +1.473, p = 0.021

BRAF-V600E activates MEK as a monomer, bypassing the RAF1-BRAF heterodimerization step required in BRAF-WT cells. In BRAF-WT signaling, CRAF (RAF1) is the primary RAF species that dimerizes with BRAF to activate MEK. In BRAF-V600E cells, the monomer is constitutively active -- RAF1 is no longer needed for MEK activation.

This is a parallel bypass to the SOS1 story: the oncogenic mutation creates a bypass that reduces dependency on the upstream or parallel component. The structural insight: mutations that create constitutive activation consistently reduce the genetic essentiality of the activating machinery upstream and in parallel -- while increasing essentiality of the driver itself.

BRAF itself trends toward BRAF-V600E essentiality in the SKCM data (d=-0.65, p=0.11, n=17/9 -- underpowered), consistent with BRAF addiction in V600E cells.

SOS1 shows no differential in BRAF-V600E SKCM (|delta|=0.079, p=0.63) -- H3 CONFIRMED. The SOS1 signal is KRAS-specific, not a general MAPK-hyperactivation effect.

What This Means for the Drug Discovery Framework

The DAMPER-Essentiality Decoupling principle changes how we interpret structural classifications in the drug discovery context:

For genetic essentiality screens (CRISPR, RNAi): Focus on ANCHOR genes -- these are the targets where knockout kills the cancer cell. KRAS, EGFR, driver genes generally.

For pharmacological inhibitors (small molecules, allosteric inhibitors): DAMPERs are better targets. The structural DAMPER classification reflects a gene that is maximally network-disruptive when its allosteric function is perturbed. BI-3406/SOS1, MEK inhibitors in ODS-regime cancers -- these are DAMPER inhibitors.

Combination logic: The Drug Structural Complementarity principle (D4, pre-registered) showed that pairing drugs with structurally complementary targets (one DAMPER + one ANCHOR) produces greater Bliss synergy than DAMPER+DAMPER or ANCHOR+ANCHOR (p=0.0003, NCI-ALMANAC). D22 adds mechanistic clarity to why: DAMPERs disrupt the allosteric network; ANCHORs cut the survival dependency. A combination that does both simultaneously is attacking two orthogonal vulnerability axes.

All pre-registration files, with timestamps and immutable hashes, are at github.com/vladi160/preregistrations. Pre-registration hash: 07da797.