Pooled rare variants

pooled_rare() serves cpg0032-pooled-rare, already aggregated to one profile per barcode. There is no plate, well or cell count in obs: this is the example of a pre-aggregated input, the shape a screen takes once wells have been collapsed to perturbations upstream. The 290 barcodes cover 132 genes, each barcode either a wild-type gene or one specific coding variant of it, so the genes come as allele series.

Use it if your own data is already aggregated to one profile per perturbation, with no well or cell counts, or is an allele series (a wild type plus coding variants of the same gene). No tutorial is built on it; the profile-space tools apply directly, since the aggregation is already done.

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

import mantispy as mt

adata = mt.ds.pooled_rare()
adata

AnnData object with n_obs × n_vars = 290 × 4417
    obs: 'Metadata_Foci_Barcode_MatchedTo_GeneCode', 'Metadata_Perturbation', 'Metadata_Perturbation_Type', 'Metadata_Gene', 'Metadata_Allele'
    var: 'object', 'feature_group', 'feature', 'channel', 'scale', 'angle', 'gray_levels', 'radial_bin', 'params', 'is_feature'
    uns: 'mantispy'
    layers: None (.X)

An allele series

Metadata_Gene is the gene and Metadata_Allele the full variant label, so the rows of one gene are its wild type and its variants.

obs = adata.obs
obs.loc[obs["Metadata_Gene"] == "ACTB", ["Metadata_Gene", "Metadata_Allele", "Metadata_Perturbation"]]

	Metadata_Gene	Metadata_Allele	Metadata_Perturbation
0	ACTB	ACTB	ACTB
1	ACTB	ACTB E364K	ACTB E364K
2	ACTB	ACTB R183W	ACTB R183W

Barcodes in profile space

On the first two principal components each point is a barcode. Colouring the wild-type genes apart from the variant alleles shows how far a coding change moved the profile from its reference.

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["allele"] = adata.obs["Metadata_Allele"].to_numpy()
is_variant = adata.obs["Metadata_Allele"].astype(str) != adata.obs["Metadata_Gene"].astype(str)
coords["kind"] = np.where(is_variant.to_numpy(), "variant", "wild type")
fig = px.scatter(
    coords,
    x="PC1",
    y="PC2",
    color="kind",
    hover_name="allele",
    title="Pooled rare variants in profile space",
)
fig.show()