JUMP CRISPR arm

jump_crispr() serves the CRISPR arm of cpg0016-jump-assembled [Chandrasekaran et al., 2023], 51,185 wells of knockouts in U2OS cells, corrected for well position and cell count and feature-selected by jump-profiling-recipe. Each gene is knocked out by several guides, and the guides are its replicates, so this is an arrayed, well-level CRISPR screen, the counterpart to the single-cell optical pooled screens scallops_arv471() and cp_posh().

Use it if your own data is an arrayed (well-level) genetic screen, knockout or otherwise, where several reagents share a target and act as its replicates. CRISPR knockouts runs the knockout workflow on it, and Published profiles and JUMP loads it.

import numpy as np
import pandas as pd
import plotly.express as px
import scanpy as sc

import mantispy as mt

The well-level object is large, so this page uses the gene-level consensus, one modz profile per gene over its guides.

adata = mt.ds.jump_crispr(aggregated=True)
adata

AnnData object with n_obs × n_vars = 7974 × 595
    obs: 'Metadata_Gene', 'Metadata_ReplicateCount', 'Metadata_Source', 'Metadata_JCP2022', 'Metadata_Control_Type', 'Metadata_Perturbation', 'Metadata_Perturbation_Type', 'Metadata_Control', 'Metadata_ChromosomeArm'
    var: 'object', 'feature_group', 'feature', 'channel', 'scale', 'angle', 'gray_levels', 'radial_bin', 'params', 'is_feature'
    uns: 'mantispy'
    layers: None (.X)

Controls against treatments

Metadata_Control_Type labels each gene consensus a negative control (negcon), a positive control (poscon) or a treatment (trt). On the first two principal components the controls should sit apart from the bulk of the genes.

scaled = adata.copy()
sc.pp.scale(scaled)
scaled.X = np.nan_to_num(scaled.X)
sc.pp.pca(scaled, n_comps=10)
coords = pd.DataFrame(scaled.obsm["X_pca"][:, :2], columns=["PC1", "PC2"])
coords["gene"] = adata.obs["Metadata_Gene"].to_numpy()
coords["control type"] = adata.obs["Metadata_Control_Type"].astype(str).to_numpy()
fig = px.scatter(
    coords,
    x="PC1",
    y="PC2",
    color="control type",
    hover_name="gene",
    title="JUMP CRISPR genes in profile space",
    opacity=0.5,
)
fig.show()