Bio Genome Engineering Grna Design

Version Compatibility

Reference examples tested with: BioPython 1.83+, CRISPOR 5.0+ (web/CLI).

Before using code patterns, verify installed versions match. If versions differ:

• Python: pip show <package> then help(module.function) to check signatures
• CLI: <tool> --version then <tool> --help to confirm flags

If code throws ImportError, AttributeError, or TypeError, introspect the installed package and adapt the example to match the actual API rather than retrying.

Output depends on inputs more than tool versions: on-target scores are model-specific and not interchangeable (a 0.7 Azimuth score is not a 0.7 CRISPRscan score), and the valid model is set by how the guide is delivered/transcribed, not by preference. Record the nuclease, the delivery context (U6/lentiviral vs in-vitro T7/RNP), and the reference genome build used for any off-target step.

Guide RNA Design

"Design guide RNAs to knock out my gene" -> Establish the delivery context, scan the target for the nuclease's PAM on both strands, drop guides that fail hard filters, rank survivors with the context-valid on-target model, choose the cut site by exon/transcript biology, and prefer guides whose predicted indel spectrum is frameshift-rich.

• Python: enumerate PAMs and apply hard filters with Bio.Seq + re; compute a Bae-style microhomology out-of-frame score
• CLI/web: crispor.py <genome> in.fa out.tsv aggregates the context-appropriate on-target score + off-target nomination per genome
• Web/code: inDelphi / FORECasT / Lindel for the full repair-outcome distribution

The Single Most Important Modern Insight -- a guide produces a reproducible indel distribution, not "a cut", and knockout success is a property of that distribution

Two facts that naive design ignores and that pass review constantly:

1. On-target efficiency scores are weak, context-locked predictors. Rule Set 2, CRISPRscan, and DeepCas9 scores correlate with measured cutting at only Spearman ~0.4 across realistic contexts (~0.7 is the ceiling even within one matched context; the same guides re-tested in another cell line correlate ~0.37-0.48). Each was trained on one assay -- U6-Pol-III lentiviral vs in-vitro T7 vs RNP -- and does not transfer across nuclease, delivery, promoter, cell type, or temperature (Haeussler 2016). Using CRISPRscan (T7/zebrafish-trained) to rank guides for a U6 lentiviral screen is a category error. Rank to shortlist, then design 3-6 guides and validate -- never trust the rank as truth.

2. Efficient editing is not knockout. A cut yields a characteristic, reproducible set of indels (Shen 2018; Allen 2019; Chen 2019); roughly 1/3 of indels are in-frame, so a 95%-efficient guide can still leave functional protein. Worse, even a confirmed frameshift may not eliminate protein -- translation reinitiation, exon skipping, NMD escape, and transcriptional adaptation rescue ~1/3 of verified knockouts (Smits 2019; Mou 2017; El-Brolosy 2019). So the modern question is "which guide, at which site, produces a high out-of-frame fraction in an NMD-competent, constitutive transcript region?" -- couple an outcome model to exon biology, not just an efficiency score. Verify the knockout at the protein level.

On-Target Score Taxonomy -- each model is valid for ONE context

Treat any score as a rank-and-shortlist signal (Spearman ~0.4 across context), never an oracle.

Nuclease & PAM Taxonomy -- expanding PAM range trades away activity/specificity

Default to WT-SpCas9-NGG; escalate to NG/xCas9/SpRY only when no acceptable NGG sits in the required window, and expect to validate harder (the valid on-target score and the off-target burden both change).

Decision Tree by Scenario

Enumerate and Filter Candidate Guides

Goal: Return valid candidate spacers for a target, on both strands, dropping guides that cannot work in the chosen delivery context.

Approach: Scan both strands for the nuclease's PAM, extract the protospacer upstream (Cas9) or downstream (Cas12a) of each PAM, and apply hard filters -- reject TTTT (Pol-III terminator) for U6/H1 expression, flag a missing 5' G for U6 (prepend a G rather than replace the first base), and note GC outside ~40-70% as a soft penalty. Ranking comes from the context-valid model (route to CRISPOR), not from a hand-rolled score.

from Bio.Seq import Seq
import re

GC_MIN, GC_MAX = 0.40, 0.70   # outside this band on-target activity falls off (Doench 2014); soft penalty

def find_guides(sequence, pam='NGG', guide_length=20):
    '''Enumerate SpCas9 (NGG) spacers on both strands; spacer is 5' of the PAM.'''
    seq = sequence.upper()
    guides = []
    for m in re.finditer(r'(?=([ACGT]GG))', seq):
        pos = m.start()
        if pos >= guide_length:
            guides.append({'spacer': seq[pos - guide_length:pos], 'pam': seq[pos:pos + 3],
                           'cut': pos - 3, 'strand': '+'})   # SpCas9 cuts ~3 bp 5' of the PAM
    rc = str(Seq(seq).reverse_complement())
    n = len(seq)
    for m in re.finditer(r'(?=([ACGT]GG))', rc):
        pos = m.start()
        if pos >= guide_length:
            guides.append({'spacer': rc[pos - guide_length:pos], 'pam': rc[pos:pos + 3],
                           'cut': n - (pos - 3), 'strand': '-'})
    return guides

def passes_u6_filters(spacer):
    '''Hard filters for U6/H1 Pol-III expression (NOT applicable to in-vitro T7/RNP).'''
    gc = sum(c in 'GC' for c in spacer) / len(spacer)
    return 'TTTT' not in spacer and GC_MIN <= gc <= GC_MAX   # TTTT terminates Pol III

Rank On-Target Activity in the Valid Context

Goal: Shortlist guides by predicted cutting using the model that matches the delivery context.

Approach: Do NOT hand-roll a scoring matrix. Route to CRISPOR, which selects the context-appropriate score (Rule Set 2/Azimuth for U6/lentiviral, CRISPRscan for T7/embryo) per the Haeussler 2016 logic and also nominates off-targets against the chosen genome. Treat the returned score as a shortlist signal, then carry 3-6 candidates forward.

# CRISPOR: aggregates the context-valid on-target score + off-target nomination per genome
crispor.py hg38 target.fa guides.tsv --maxOcc 60000
# columns include the on-target score (context-selected) and off-target counts/specificity

Choose the Cut Site by Exon Biology (the under-used lever)

KO success is mostly won here, and pure efficiency ranking fails:

• Target an early, constitutive coding exon (present in all protein-coding isoforms) -- but not the start-ATG region (downstream reinitiation can rescue an N-terminal truncation).
• Avoid the last exon and the last ~50 nt of the penultimate exon -- PTCs there escape NMD, leaving a stable, possibly-functional truncated protein.
• Keep the cut away from splice donor/acceptor sites unless splice disruption is the goal -- indels there cause exon skipping that can restore frame (Mou 2017).
• Confirm the exon is constitutive in the cell type of interest (an exon spliced out of the dominant isoform is a silent failure), and screen for SNPs under the protospacer/PAM in the actual background (mismatch/PAM loss -> allele dropout).
• For ruthless KO / screens: tile the conserved functional domain (Shi 2015), not the gene start.

Predict the Editing Outcome (frameshift fraction decides KO)

Goal: Prefer guides whose predicted indel spectrum is frameshift-rich (and, for a single-genotype line, dominated by one outcome).

Approach: Cas9 repair outcomes are predictable from the ~30 bp of local sequence flanking the cut. The cheap, no-ML signal is the Bae 2014 microhomology out-of-frame score: enumerate microhomology pairs flanking the cut, weight each predicted MMEJ deletion, and report the fraction whose length is not a multiple of 3. For a full genotype distribution use inDelphi (Shen 2018), FORECasT (Allen 2019), or Lindel (Chen 2019). Rank by (editing efficiency) x (out-of-frame fraction) -- a 70%-efficient guide with frameshift fraction 0.9 beats a 90%-efficient guide at 0.5. (See examples/grna_design.py for a runnable Bae-style out-of-frame implementation.)

Per-Method Failure Modes

"We used the top-ranked guide" with no validation

Trigger: sorting by on-target score and taking #1. Mechanism: scores are Spearman ~0.4 across context. Symptom: confident ranking, poor empirical hit rate. Fix: design 3-6 guides per gene and validate; treat the score as triage.