pen-stack 6.10.1__tar.gz → 6.10.2__tar.gz

{pen_stack-6.10.1 → pen_stack-6.10.2}/CHANGELOG.md

@@ -3,6 +3,31 @@
 All notable changes to PEN-STACK are documented here. This file follows
 [Keep a Changelog](https://keepachangelog.com/) and the program's phase structure.
 
+## [6.10.2] - 2026-06-20 - Chromatin context: controlled validation (honest weak/inconsistent result)
+
+**PATCH — validate the chromatin axis for real, then scope it to what the data supports.** v6.10.1 wired a real
+Stage B accessibility read but had not *demonstrated* it carries signal. v6.10.2 runs a controlled experiment and
+reports the result honestly — it is **not** a clean win.
+
+### Added
+- **A controlled chromatin validation** (`benchmarks/offtarget/chromatin_validation.json`,
+  `scripts/offtarget_chromatin_validation.py`). Off-target protospacers mapped to hg38 (GRCh38.fa, exact match both
+  strands, 98.5% mapped); AUROC of the Stage B K562 ATAC/DNase accessibility for active-vs-inactive off-targets per
+  assay, with **in-vitro assays as a NEGATIVE control** (cell-free → no chromatin). Result: the in-vitro controls hug
+  0.5 (method sound — no spurious signal); the canonical cell-based assay **GUIDE-seq is a textbook modest positive
+  (AUROC 0.58, CI [0.550, 0.613])** even with a cross-cell-type K562 proxy; but the second cell-based assay **TTISS
+  reverses (0.346)**. **Verdict: WEAK / INCONSISTENT — chromatin is NOT a validated quantitative axis on this data**
+  (the cross-cell-type proxy is the likely cause; a cell-type-matched accessibility track would settle it — deferred).
+
+### Changed
+- The chromatin-accessibility modifier is now an explicit **annotation only**, labelled `validated=false`: it reads
+  the real Stage B track (or a caller-supplied scalar) and notes the qualitative Lazzarotto-2020 direction, but it
+  **does not change the numeric off-target risk score**. `offtarget_data.CHROMATIN_VALIDATION` carries the result.
+- Corrected the v6.10.1 disclosures-ledger entry that had marked chromatin "FIXED" — that was too strong;
+  it is now disclosed as *tested → weak/inconsistent → documented annotation, not a validated axis*.
+
+This is an honest negative-ish result reported in full rather than dropped or overstated.
+
 ## [6.10.1] - 2026-06-20 - WS-OFFTARGET rigor pass + retroactive honesty audit (v6.7–v6.9)
 
 **PATCH — complete the Stage E plan and disclose every remaining deviation.** Closes the v6.10.0 gaps and adds a

{pen_stack-6.10.1 → pen_stack-6.10.2}/CITATION.cff

@@ -1,7 +1,7 @@
 cff-version: 1.2.0
 message: "If you use PEN-STACK, please cite it as below."
 title: "PEN-STACK: open infrastructure for genome writing"
-version: 6.10.1
+version: 6.10.2
 date-released: 2026-06-20
 authors:
   - family-names: "Mahaboob Ali"

{pen_stack-6.10.1 → pen_stack-6.10.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: pen-stack
-Version: 6.10.1
+Version: 6.10.2
 Summary: Open infrastructure for genome writing: the Writable Genome atlas, the Writer Atlas, and the Write Planner.
 Author-email: Anees Ahmed Mahaboob Ali <ahmedaneesm@gmail.com>
 License: MIT
@@ -90,7 +90,7 @@ every design against rule-grounded mechanism, reports calibrated confidence, cit
 [![codecov](https://codecov.io/gh/ahmedanees-m/pen-stack/branch/main/graph/badge.svg)](https://codecov.io/gh/ahmedanees-m/pen-stack)
 [![License: MIT](https://img.shields.io/badge/License-MIT-informational.svg)](LICENSE)
 [![Python 3.11+](https://img.shields.io/badge/python-3.11%2B-blue.svg)](https://www.python.org/)
-[![Version](https://img.shields.io/badge/version-6.10.1-blue.svg)](CHANGELOG.md)
+[![Version](https://img.shields.io/badge/version-6.10.2-blue.svg)](CHANGELOG.md)
 [![Status](https://img.shields.io/badge/status-1.0%20First%20Stable-success.svg)](docs/STABILITY.md)
 [![Tests](https://img.shields.io/badge/tests-378%20passing-success.svg)](tests/)
 [![Lint: ruff](https://img.shields.io/badge/lint-ruff-purple.svg)](https://github.com/astral-sh/ruff)

{pen_stack-6.10.1 → pen_stack-6.10.2}/README.md

@@ -15,7 +15,7 @@ every design against rule-grounded mechanism, reports calibrated confidence, cit
 [![codecov](https://codecov.io/gh/ahmedanees-m/pen-stack/branch/main/graph/badge.svg)](https://codecov.io/gh/ahmedanees-m/pen-stack)
 [![License: MIT](https://img.shields.io/badge/License-MIT-informational.svg)](LICENSE)
 [![Python 3.11+](https://img.shields.io/badge/python-3.11%2B-blue.svg)](https://www.python.org/)
-[![Version](https://img.shields.io/badge/version-6.10.1-blue.svg)](CHANGELOG.md)
+[![Version](https://img.shields.io/badge/version-6.10.2-blue.svg)](CHANGELOG.md)
 [![Status](https://img.shields.io/badge/status-1.0%20First%20Stable-success.svg)](docs/STABILITY.md)
 [![Tests](https://img.shields.io/badge/tests-378%20passing-success.svg)](tests/)
 [![Lint: ruff](https://img.shields.io/badge/lint-ruff-purple.svg)](https://github.com/astral-sh/ruff)

{pen_stack-6.10.1 → pen_stack-6.10.2}/docs/DEVIATIONS_AND_DISCLOSURES.md

@@ -16,7 +16,7 @@ applicable (e.g. single contributor).
 |---|---|---|
 | Nuclease predictor named in plan (CCLMoff / CRISMER) vs shipped (CRISOT) | **DISCLOSED** | We use **CRISOT-Score** (Chen 2023, CC-BY-NC, CPU), not the plan's CCLMoff/CRISMER. Justified: **CRISMER ships no license** (cannot be redistributed/wrapped); **CCLMoff** is a GPU RNA-language-model stack AND was trained on these very assays (evaluating it on them would be **leakage**), whereas **CRISOT-Score is MD-physics, assay-agnostic → a leakage-clean held-out evaluation**. CRISOT is the more rigorous and license-appropriate choice. A genuine substitution, disclosed. |
 | Assay coverage (plan listed GUIDE/CIRCLE/CHANGE/SITE/SURRO/TTISS/Digenome-seq) | **FIXED (v6.10.1)** | v6.10.0 used GUIDE-seq + CIRCLE-seq only. v6.10.1 adds **CHANGE-seq + SITE-seq** (independent broad guide panels) — CRISOT beats homology on **all four** (per-guide bootstrap CI excludes 0). SURRO-seq is targeted/biased (kept as an orthogonal reference, not genome-wide truth); TTISS/Digenome-seq remain **DISCLOSED** as not-yet-folded-in. |
-| "Chromatin-aware via the Stage B feature store" | **FIXED (v6.10.1)** | v6.10.0 had only a caller-supplied accessibility hook and the docs said "chromatin-aware" (overclaim). v6.10.1 adds a **real** `locus_accessibility(chrom, bin, ct)` that reads the Stage B `chromatin_{ct}.parquet` ATAC/DNase track when present, **abstains** when absent (bare wheel / CI / current deployed atlas do not ship the raw track), and accepts a caller-supplied scalar otherwise. Docs relabelled to "chromatin-accessibility modifier". |
+| "Chromatin-aware via the Stage B feature store" | **DISCLOSED (tested v6.10.2 — weak/inconsistent)** | v6.10.0 overclaimed "chromatin-aware" with only a caller-supplied hook. v6.10.1 added a **real** `locus_accessibility(chrom, bin, ct)` reading the Stage B `chromatin_{ct}.parquet` ATAC/DNase track (verified on the VM), abstaining when absent, and relabelled "chromatin-accessibility modifier". **v6.10.2 ran a controlled validation** (off-targets mapped to hg38, AUROC of K562 accessibility for active-vs-inactive per assay, in-vitro negative controls): the in-vitro controls are ~0.5 (method sound), GUIDE-seq is a textbook modest positive (**0.58**, CI excludes 0.5) but TTISS **reverses** (0.346) — the cross-cell-type K562 proxy is the likely cause. **VERDICT: weak/inconsistent → NOT a validated quantitative axis.** Chromatin stays a documented **annotation only** (does not change the numeric risk score), labelled `validated=false`. Full result: `benchmarks/offtarget/chromatin_validation.json`. The definitive refinement (cell-type-matched ENCODE accessibility) is deferred. *This corrects the v6.10.1 ledger entry, which had marked it "FIXED" — that was too strong.* |
 | Held-out **guide** AND **locus** splits | **DISCLOSED** | Held-out-**guide** is implemented; a genomic-coordinate **locus** split is **not possible** — the harmonized assay data ships sequences, not coordinates. Instead we add **cross-assay generalization** (the assay-agnostic CRISOT-Score evaluated on GUIDE/CIRCLE canonical guides + CHANGE/SITE independent panels). This is the leakage-clean substitute and is stated in `benchmarks/offtarget/split.json`. |
 | Learned **integrase** off-target scorer (plan: "learned on HIDE/Cryptic-seq") | **DISCLOSED** | Shipped as a documented **pseudo-attB cryptic scan** on the real Bxb1 attB core, not a learned model — Cryptic-seq/HIDE-seq are recent preprints and IntQuery has **no public weights**, so a real learned integrase predictor is not groundable. Flagged extrapolative; IntQuery cited as paper-only. |
 | Bench "joins the Challenge" | **FIXED (v6.10.1)** | An `offtarget` nomination task is added to the Genome-Writing Challenge (`benchmarks/genome_writing_challenge/harness.py`) — non-circular (label = wet-lab Active call), data-gated on the fixture. |

{pen_stack-6.10.1 → pen_stack-6.10.2}/docs/cards/offtarget_data.md

@@ -41,13 +41,29 @@ evaluation on every assay.
 The learned predictor beats the homology baseline on **all four** assays (per-guide bootstrap CI excludes 0),
 including two independent broad guide panels (cross-assay generalization).
 
-## Chromatin-accessibility modifier (honest scope)
+## Chromatin-accessibility modifier (honest scope — tested, NOT validated)
 The nominator applies a documented chromatin modifier (open chromatin raises realized off-target activity;
 Lazzarotto 2020). It reads the **real Stage B accessibility track** (`phase_1/features/chromatin_{ct}.parquet`,
-ATAC/DNase) when a candidate's genomic locus + cell type are supplied AND the feature store is present, or accepts a
-caller-supplied scalar; it **abstains** otherwise. The bare wheel / CI / current deployed atlas do **not** ship the
-raw accessibility track, so the modifier is inactive there — this is a documented, honestly-bounded integration, not
-an autonomous "chromatin-aware" guarantee.
+ATAC/DNase) when a candidate's genomic locus + cell type are supplied AND the feature store is present (verified on
+the VM), or accepts a caller-supplied scalar; it **abstains** otherwise (the bare wheel / CI / deployed atlas do not
+ship the raw track).
+
+**v6.10.2 controlled validation** (`benchmarks/offtarget/chromatin_validation.json`): off-targets mapped to hg38
+(98.5%), AUROC of K562 accessibility for active-vs-inactive off-targets per assay, with in-vitro assays as a
+NEGATIVE control. Result — **weak / inconsistent**:
+
+| Assay | modality | accessibility AUROC | 95% CI |
+|---|---|---|---|
+| GUIDE-seq | cell-based | **0.58** | [0.550, 0.613] — excludes 0.5 |
+| TTISS | cell-based | 0.346 | [0.322, 0.368] — *reversed* |
+| CHANGE-seq | in-vitro control | 0.496 | [0.480, 0.513] — null ✓ |
+| CIRCLE-seq | in-vitro control | 0.519 | [0.501, 0.535] — ~null ✓ |
+| SITE-seq | in-vitro control | 0.469 | [0.452, 0.489] — ~null ✓ |
+
+The in-vitro controls hug 0.5 (the method has no spurious signal); the canonical cell-based assay (GUIDE-seq) shows
+the textbook modest positive even with a **cross-cell-type** K562 proxy, but TTISS reverses. **Verdict: NOT a
+validated quantitative axis** — chromatin is a documented **annotation only** (it does not change the numeric risk
+score), labelled `validated=false`. A cell-type-matched accessibility track is the definitive refinement (deferred).
 
 ## Risk calibration (grounded)
 The nomination risk band IS the empirical fraction of candidates at *k* mismatches that were validated-active

{pen_stack-6.10.1 → pen_stack-6.10.2}/docs/offtarget.md

@@ -15,7 +15,11 @@ A nominated off-target is a candidate, and every result ships with the empirical
   modifier** (open chromatin raises realized off-target activity; Lazzarotto 2020). The modifier reads the **real
   Stage B accessibility track** (`phase_1/features/chromatin_{ct}.parquet`) when a candidate's genomic locus + cell
   type are supplied and the store is present, accepts a caller-supplied scalar otherwise, and **abstains** when
-  neither is available (the bare wheel / current deployed atlas do not ship the raw track).
+  neither is available (the bare wheel / current deployed atlas do not ship the raw track). A **v6.10.2 controlled
+  validation** (off-targets mapped to hg38; AUROC of accessibility for active-vs-inactive off-targets, in-vitro
+  negative controls) found the signal **weak/inconsistent** (GUIDE-seq 0.58 but TTISS 0.346; in-vitro controls
+  ~null) — so chromatin is an **annotation only** (`validated=false`), it does **not** change the numeric risk
+  score. Full result: `benchmarks/offtarget/chromatin_validation.json`.
 - **Serine integrase (Bxb1):** a cryptic **pseudo-attB** scan that seeds on the *real documented* Bxb1 attB core
   (`GCGGTCTC`, central GT; FlyBase FBto0000359, Ghosh 2003) and reports candidate cryptic sites by arm mismatches.
 - **Bridge recombinase:** delegates to the existing Perry-DMS pseudosite engine (`pen_stack.bridge.offtarget`).

{pen_stack-6.10.1 → pen_stack-6.10.2}/pen_stack/__init__.py

@@ -1,2 +1,2 @@
 """PEN-STACK v3.0 - open infrastructure for genome writing."""
-__version__ = "6.10.1"
+__version__ = "6.10.2"

{pen_stack-6.10.1 → pen_stack-6.10.2}/pen_stack/wgenome/offtarget_data.py

@@ -57,6 +57,27 @@ MISMATCH_ACTIVE_FRACTION = {
     "siteseq": {0: 1.0, 1: 1.0, 2: 1.0, 3: 0.67188, 4: 0.2491, 5: 0.03554, 6: 0.00478},
 }
 
+# ---- chromatin-accessibility validation (v6.10.2): a CONTROLLED test of whether Stage B accessibility predicts
+# off-target activity (Lazzarotto 2020). VERDICT: WEAK / INCONSISTENT -> NOT a validated quantitative axis.
+# (off-targets mapped to hg38; AUROC of K562 accessibility for active-vs-inactive per assay; full result in
+# benchmarks/offtarget/chromatin_validation.json). The in-vitro controls are ~null (method sound); GUIDE-seq is a
+# textbook modest positive (0.58) but TTISS reverses (0.346) — the cross-cell-type K562 proxy is the likely cause.
+CHROMATIN_VALIDATION = {
+    "verdict": "weak/inconsistent — NOT a validated quantitative axis on this data",
+    "auroc_accessibility_for_activity": {
+        "guideseq": {"modality": "cell-based", "auroc": 0.58, "ci95": [0.550, 0.613]},
+        "ttiss": {"modality": "cell-based", "auroc": 0.346, "ci95": [0.322, 0.368]},
+        "changeseq": {"modality": "in_vitro_control", "auroc": 0.496, "ci95": [0.480, 0.513]},
+        "circleseq": {"modality": "in_vitro_control", "auroc": 0.519, "ci95": [0.501, 0.535]},
+        "siteseq": {"modality": "in_vitro_control", "auroc": 0.469, "ci95": [0.452, 0.489]},
+    },
+    "validated": False,
+    "note": "documented biological effect (Lazzarotto 2020); on our cross-cell-type data the signal is weak and "
+            "inconsistent (GUIDE-seq 0.58 supports it, TTISS 0.346 contradicts, in-vitro controls ~null). Used as a "
+            "qualitative annotation only — it does NOT change the numeric risk score. A cell-type-matched "
+            "accessibility track would be needed to settle it.",
+}
+
 # ---- Off-Target-Bench headline (REAL full-data result; per-guide AUPRC, held-out-guide bootstrap CI) -----
 # CRISOT-Score is the MD-physics, assay-AGNOSTIC scorer (not fit on these labels) -> a leakage-clean held-out
 # evaluation on every assay. GUIDE/CIRCLE-seq use canonical guides; CHANGE/SITE-seq use INDEPENDENT broad guide

{pen_stack-6.10.1 → pen_stack-6.10.2}/pen_stack/wgenome/offtarget_predict.py

@@ -98,26 +98,28 @@ def locus_accessibility(chrom: str, bin_idx: int, ct: str = "k562") -> dict | No
 
 
 def _chromatin_modifier(accessibility: float | None = None, locus_acc: dict | None = None) -> dict | None:
-    """Documented chromatin-accessibility modifier: open chromatin raises realized off-target activity
-    (Lazzarotto et al. 2020, CHANGE-seq, 10.1038/s41587-020-0555-7). Uses the REAL Stage B accessibility
-    (`locus_acc`) when available, else a caller-supplied 0..1 scalar; abstains (None) when neither is given.
-    Bounded/qualitative — never a fabricated magnitude."""
+    """Documented chromatin-accessibility ANNOTATION (NOT a validated quantitative axis): open chromatin raises
+    realized off-target activity (Lazzarotto et al. 2020, CHANGE-seq, 10.1038/s41587-020-0555-7). Uses the REAL
+    Stage B accessibility (`locus_acc`) when available, else a caller-supplied 0..1 scalar; abstains (None) when
+    neither is given. It is QUALITATIVE and does NOT change the numeric risk score — a controlled test on our data
+    found the signal weak/inconsistent (see `offtarget_data.CHROMATIN_VALIDATION`)."""
+    val = {"validated_on_our_data": False, "validation": "weak/inconsistent (GUIDE-seq AUROC 0.58, TTISS 0.346, "
+           "in-vitro controls ~null; cross-cell-type proxy) — annotation only, does not change the risk score",
+           "doi": "10.1038/s41587-020-0555-7"}
     if locus_acc is not None:
         raises = bool(locus_acc.get("accessible"))
-        return {"raises_realized_risk": raises, "source": "stage_b_track",
+        return {**val, "raises_realized_risk": raises, "source": "stage_b_track",
                 "accessibility_signal": locus_acc.get("accessibility_signal"), "cell_type": locus_acc.get("cell_type"),
-                "effect": ("accessible (open) chromatin in this cell type -> realized off-target activity is HIGHER "
-                           "for the same sequence match (qualitative; Lazzarotto 2020)" if raises
-                           else "inaccessible (closed) chromatin -> realized off-target activity is lower"),
-                "doi": "10.1038/s41587-020-0555-7"}
+                "effect": ("accessible (open) chromatin in this cell type -> realized off-target activity tends "
+                           "HIGHER for the same sequence match (qualitative; Lazzarotto 2020)" if raises
+                           else "inaccessible (closed) chromatin -> realized off-target activity tends lower")}
     if accessibility is None:
         return None
     acc = max(0.0, min(1.0, float(accessibility)))
     raises = acc >= 0.5
-    return {"accessibility": round(acc, 3), "raises_realized_risk": raises, "source": "caller_supplied",
-            "effect": ("open/accessible chromatin -> realized off-target activity is HIGHER (qualitative; "
-                       "Lazzarotto 2020)" if raises else "closed/inaccessible chromatin -> realized risk lower"),
-            "doi": "10.1038/s41587-020-0555-7"}
+    return {**val, "accessibility": round(acc, 3), "raises_realized_risk": raises, "source": "caller_supplied",
+            "effect": ("open/accessible chromatin -> realized off-target activity tends HIGHER (qualitative; "
+                       "Lazzarotto 2020)" if raises else "closed/inaccessible chromatin -> realized risk tends lower")}
 
 
 def nominate_nuclease(guide: str, candidate_sites: list[str] | None, accessibility: list[float] | None = None,

{pen_stack-6.10.1 → pen_stack-6.10.2}/pen_stack.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: pen-stack
-Version: 6.10.1
+Version: 6.10.2
 Summary: Open infrastructure for genome writing: the Writable Genome atlas, the Writer Atlas, and the Write Planner.
 Author-email: Anees Ahmed Mahaboob Ali <ahmedaneesm@gmail.com>
 License: MIT
@@ -90,7 +90,7 @@ every design against rule-grounded mechanism, reports calibrated confidence, cit
 [![codecov](https://codecov.io/gh/ahmedanees-m/pen-stack/branch/main/graph/badge.svg)](https://codecov.io/gh/ahmedanees-m/pen-stack)
 [![License: MIT](https://img.shields.io/badge/License-MIT-informational.svg)](LICENSE)
 [![Python 3.11+](https://img.shields.io/badge/python-3.11%2B-blue.svg)](https://www.python.org/)
-[![Version](https://img.shields.io/badge/version-6.10.1-blue.svg)](CHANGELOG.md)
+[![Version](https://img.shields.io/badge/version-6.10.2-blue.svg)](CHANGELOG.md)
 [![Status](https://img.shields.io/badge/status-1.0%20First%20Stable-success.svg)](docs/STABILITY.md)
 [![Tests](https://img.shields.io/badge/tests-378%20passing-success.svg)](tests/)
 [![Lint: ruff](https://img.shields.io/badge/lint-ruff-purple.svg)](https://github.com/astral-sh/ruff)

{pen_stack-6.10.1 → pen_stack-6.10.2}/pen_stack.egg-info/SOURCES.txt

@@ -480,6 +480,7 @@ prereg/ws_writer.yaml
 prereg/ws_wv.yaml
 scripts/calibrate_immune_axes.py
 scripts/fetch_licensed_sources.py
+scripts/offtarget_chromatin_validation.py
 scripts/p1_build_atlas.py
 scripts/p1_build_durability.py
 scripts/p1_build_position_effect.py

{pen_stack-6.10.1 → pen_stack-6.10.2}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "pen-stack"
-version = "6.10.1"
+version = "6.10.2"
 description = "Open infrastructure for genome writing: the Writable Genome atlas, the Writer Atlas, and the Write Planner."
 readme = "README.md"
 requires-python = ">=3.11"

pen_stack-6.10.2/scripts/offtarget_chromatin_validation.py

@@ -0,0 +1,115 @@
+"""Reproducible analysis (v6.10.2): does chromatin accessibility predict off-target activity? (Lazzarotto 2020).
+
+Run on the VM (needs hg38 GRCh38.fa, the Stage B chromatin_{ct}.parquet, and the harmonized assay CSVs). Maps each
+off-target protospacer to hg38 coordinates by exact match on both strands (grep -F, one genome pass), intersects with
+the Stage B K562 ATAC/DNase accessibility at 1 kb bins, and computes the AUROC of accessibility for active-vs-inactive
+off-targets PER ASSAY. Cell-based assays (GUIDE-seq, TTISS) are the test; in-vitro assays (CHANGE/CIRCLE/SITE-seq) are
+the NEGATIVE control (cell-free -> no chromatin -> accessibility should not discriminate). Writes the result JSON.
+
+RESULT (committed in benchmarks/offtarget/chromatin_validation.json): WEAK / INCONSISTENT — the in-vitro controls are
+~0.5 (method sound), GUIDE-seq is a textbook modest positive (0.58) but TTISS reverses (0.346); the cross-cell-type
+K562 accessibility proxy is the likely cause. Chromatin is therefore a documented annotation, NOT a validated axis.
+
+Usage (on the VM, inside a container with pandas+numpy):
+    FA=.../GRCh38.fa CHROM=.../chromatin_k562.parquet DS=.../datasets python offtarget_chromatin_validation.py
+"""
+from __future__ import annotations
+
+import collections
+import json
+import os
+import subprocess
+
+import numpy as np
+import pandas as pd
+
+FA = os.environ.get("FA", "/data/genomes/GRCh38.fa")
+CHROM = os.environ.get("CHROM", "/data/chromatin_k562.parquet")
+DS = os.environ.get("DS", "/data/datasets")
+COMP = str.maketrans("ACGT", "TGCA")
+ASSAYS = {"guideseq": "cell-based", "ttiss": "cell-based", "changeseq": "in_vitro_control",
+          "circleseq_all": "in_vitro_control", "siteseq": "in_vitro_control"}
+VALID = {f"chr{i}" for i in list(range(1, 23)) + ["X", "Y"]}
+
+
+def rc(s: str) -> str:
+    return s.translate(COMP)[::-1]
+
+
+def auroc(y, s):
+    y = np.asarray(y)
+    s = np.asarray(s, float)
+    n1, n0 = int(y.sum()), int((1 - y).sum())
+    if n1 == 0 or n0 == 0:
+        return None, n1, n0
+    order = np.argsort(s)
+    ranks = np.empty(len(s))
+    ranks[order] = np.arange(1, len(s) + 1)
+    d = collections.defaultdict(list)
+    for i, v in enumerate(s):
+        d[v].append(i)
+    for _v, idxs in d.items():
+        rr = np.mean([ranks[i] for i in idxs])
+        for i in idxs:
+            ranks[i] = rr
+    return float((ranks[y == 1].sum() - n1 * (n1 + 1) / 2) / (n1 * n0)), n1, n0
+
+
+def main() -> dict:
+    rng = np.random.RandomState(7)
+    frames = []
+    for tag in ASSAYS:
+        df = pd.read_csv(f"{DS}/{tag}.csv")
+        df["assay"] = tag.replace("_all", "")
+        inact = df[df["Active"] == 0]
+        frames.append(pd.concat([df[df["Active"] == 1], inact.sample(n=min(1500, len(inact)), random_state=rng)]))
+    sub = pd.concat(frames).reset_index(drop=True)
+    sub["off23"] = sub["Off"].str[:23]
+    pat_map = {}
+    for off in sub["off23"].unique():
+        if len(off) == 23 and set(off) <= set("ACGT"):
+            pat_map[off] = off
+            pat_map[rc(off)] = off
+    open("/tmp/pats.txt", "w").write("\n".join(pat_map) + "\n")
+    hits: dict = {}
+    cur, buf = None, []
+
+    def flush(chrom, seq):
+        c = chrom if chrom and chrom.startswith("chr") else f"chr{chrom}"
+        if not seq or c not in VALID:
+            return
+        open("/tmp/chr.txt", "w").write(seq)
+        p = subprocess.run(["grep", "-boFf", "/tmp/pats.txt", "/tmp/chr.txt"], capture_output=True, text=True)
+        for ln in p.stdout.splitlines():
+            off_b, _, matched = ln.partition(":")
+            fwd = pat_map.get(matched)
+            if fwd:
+                hits.setdefault(fwd, []).append((c, int(off_b)))
+    with open(FA) as fh:
+        for line in fh:
+            if line.startswith(">"):
+                if cur:
+                    flush(cur, "".join(buf))
+                cur, buf = line[1:].split()[0], []
+            else:
+                buf.append(line.strip().upper())
+        if cur:
+            flush(cur, "".join(buf))
+    cdf = pd.read_parquet(CHROM)
+    cdf["acc"] = cdf[["atac", "dnase"]].max(axis=1)
+    acc_idx = {(c, int(b)): float(a) for c, b, a in zip(cdf["chrom"], cdf["bin"], cdf["acc"])}
+    sub["acc"] = sub["off23"].map(lambda o: max(
+        [acc_idx.get((c, p // 1000)) for c, p in hits.get(o, []) if acc_idx.get((c, p // 1000)) is not None] or [np.nan]))
+    mp = sub.dropna(subset=["acc"])
+    out = {}
+    for tag in {t.replace("_all", "") for t in ASSAYS}:
+        a = mp[mp["assay"] == tag]
+        au, n1, n0 = auroc(a["Active"].values, a["acc"].values)
+        out[tag] = {"modality": ASSAYS.get(tag, ASSAYS.get(tag + "_all")), "auroc": au if au is None else round(au, 3),
+                    "actives": n1, "inactives": n0}
+    print(json.dumps(out, indent=2))
+    return out
+
+
+if __name__ == "__main__":
+    main()