""" Tier 2 / Step 1: Resolve a CRISPick precomputed-design download URL. CRISPick (Broad Institute GPP) publishes genome-wide precomputed sgRNA designs. The bundled `references/resource/CRISPick_download_links.txt` holds 238 URLs = 119 full datasets (`.txt.gz`, 50-700 MB each) + 119 matching summary files (`.summary.txt`, 1-3 MB). File naming convention: sgRNA_design_{TAXID}_{GENOME}_{CAS}_{APPLICATION}_{ALGORITHM}_{SOURCE}_{DATE}.txt.gz This module filters those links by organism (TAXID), Cas enzyme, and application, and returns the matching dataset URL(s). It does NOT download them (too large to bundle) — the caller runs wget + gunzip, then uses select_crispick_sgrnas.py. CRITICAL enzyme note (preserved from the source guide): AsCas12a (wild-type) and enAsCas12a (enhanced) are DIFFERENT enzymes with different activity profiles. Guides designed for one may NOT work with the other. Always match the dataset to your exact Cas12a variant. """ from __future__ import annotations import os import re _HERE = os.path.dirname(os.path.abspath(__file__)) DEFAULT_LINKS = os.path.normpath( os.path.join(_HERE, "..", "references", "resource", "CRISPick_download_links.txt") ) # Common organism -> NCBI TAXID (subset present in the bundled file: 13 organisms). ORGANISM_TAXID = { "human": "9606", "h. sapiens": "9606", "homo sapiens": "9606", "mouse": "10090", "m. musculus": "10090", "mus musculus": "10090", "rat": "10116", "r. norvegicus": "10116", "dog": "9615", "c. familiaris": "9615", "cow": "9913", "b. taurus": "9913", "monkey": "9544", "macaque": "9544", "rhesus": "9544", "m. mulatta": "9544", "fly": "7227", "drosophila": "7227", "d. melanogaster": "7227", "salmon": "8030", "pig": "9823", "chicken": "9031", } # User-facing Cas name -> token used in the file. CAS_TOKEN = { "spcas9": "SpyoCas9", "spyocas9": "SpyoCas9", "s. pyogenes": "SpyoCas9", "sp": "SpyoCas9", "sacas9": "SaurCas9", "saurcas9": "SaurCas9", "s. aureus": "SaurCas9", "sa": "SaurCas9", "ascas12a": "AsCas12a", "as": "AsCas12a", "enascas12a": "enAsCas12a", "enas": "enAsCas12a", "enhanced": "enAsCas12a", } # Cas -> PAM (for reporting / sanity). CAS_PAM = { "SpyoCas9": "NGG (3')", "SaurCas9": "NNGRRT (3')", "AsCas12a": "TTTV (5')", "enAsCas12a": "TTTV (5')", } # User-facing application -> token used in the file. APP_TOKEN = { "knockout": "CRISPRko", "ko": "CRISPRko", "cut": "CRISPRko", "crisprko": "CRISPRko", "activation": "CRISPRa", "crispra": "CRISPRa", "activate": "CRISPRa", "inhibition": "CRISPRi", "crispri": "CRISPRi", "interference": "CRISPRi", "knockdown": "CRISPRi", } def _norm(s: str) -> str: return re.sub(r"\s+", " ", str(s)).strip().lower() def load_links(links_path: str = DEFAULT_LINKS, full_only: bool = True) -> list[str]: """Return the list of CRISPick URLs. By default only full datasets (.txt.gz).""" with open(links_path) as fh: links = [ln.strip() for ln in fh if ln.strip()] if full_only: links = [ln for ln in links if ln.endswith(".txt.gz")] return links def find_crispick_dataset( organism: str, cas: str = "SpCas9", application: str = "knockout", links_path: str = DEFAULT_LINKS, full_only: bool = True, ) -> dict: """ Resolve CRISPick download URL(s) for an organism / Cas enzyme / application. Parameters ---------- organism : str Organism name or NCBI TAXID (e.g. "human", "mouse", "9606"). cas : str Cas enzyme: SpCas9 (default), SaCas9, AsCas12a, enAsCas12a. application : str "knockout"/"ko" (default), "activation"/"CRISPRa", "inhibition"/"CRISPRi". links_path : str Path to the bundled links file. full_only : bool If True, restrict to full `.txt.gz` datasets (the ones you actually parse). Returns ------- dict with keys: taxid, cas_token, pam, app_token, matches (list of URLs), warning (str or None). """ taxid = organism if re.fullmatch(r"\d+", str(organism)) else ORGANISM_TAXID.get(_norm(organism)) cas_token = CAS_TOKEN.get(_norm(cas)) app_token = APP_TOKEN.get(_norm(application)) problems = [] if taxid is None: problems.append(f"Unknown organism '{organism}'. Known: {sorted(set(ORGANISM_TAXID))}.") if cas_token is None: problems.append(f"Unknown Cas '{cas}'. Known: SpCas9, SaCas9, AsCas12a, enAsCas12a.") if app_token is None: problems.append(f"Unknown application '{application}'. Known: knockout, activation, inhibition.") links = load_links(links_path, full_only=full_only) matches = [] if not problems: for url in links: fname = url.rsplit("/", 1)[-1] # TAXID is the first numeric token after 'sgRNA_design_'. m = re.match(r"sgRNA_design_(\d+)_", fname) if not m or m.group(1) != taxid: continue # Match Cas token exactly (so 'AsCas12a' does not match 'enAsCas12a' and vice versa). if not re.search(rf"_{re.escape(cas_token)}_", fname): continue if not re.search(rf"_{re.escape(app_token)}_", fname): continue matches.append(url) # Enzyme-variant warning for Cas12a. warning = None if cas_token == "AsCas12a": warning = ("You requested wild-type AsCas12a. Do NOT use an enAsCas12a dataset — the " "enhanced variant has a different activity profile and guides may not transfer.") elif cas_token == "enAsCas12a": warning = ("You requested enhanced enAsCas12a. Do NOT use a wild-type AsCas12a dataset — " "guides designed for one variant may not work with the other.") return { "taxid": taxid, "cas_token": cas_token, "pam": CAS_PAM.get(cas_token), "app_token": app_token, "matches": matches, "problems": problems, "warning": warning, } if __name__ == "__main__": import sys org = sys.argv[1] if len(sys.argv) > 1 else "human" cas = sys.argv[2] if len(sys.argv) > 2 else "SpCas9" app = sys.argv[3] if len(sys.argv) > 3 else "knockout" r = find_crispick_dataset(org, cas, app) if r["problems"]: print("PROBLEMS:") for p in r["problems"]: print(" -", p) print(f"TAXID={r['taxid']} Cas={r['cas_token']} (PAM {r['pam']}) App={r['app_token']}") if r["warning"]: print("WARNING:", r["warning"]) print(f"{len(r['matches'])} matching dataset(s):") for u in r["matches"]: print(" ", u) if r["matches"]: print("\nDownload with:\n wget '" + r["matches"][0] + "'\n gunzip sgRNA_design_*.txt.gz")