| """ |
| Copyright (C) 2022, 2023 GlaxoSmithKline plc - Mathieu Chevalley, Yusuf Roohani; |
| |
| Licensed under the Apache License, Version 2.0 (the "License"); |
| you may not use this file except in compliance with the License. |
| You may obtain a copy of the License at |
| http://www.apache.org/licenses/LICENSE-2.0 |
| Unless required by applicable law or agreed to in writing, software |
| distributed under the License is distributed on an "AS IS" BASIS, |
| WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. |
| See the License for the specific language governing permissions and |
| limitations under the License. |
| """ |
| import numpy as np |
| import pandas as pd |
| import scanpy as sc |
|
|
| def get_strong_perts(supp): |
| filtered = supp[supp['Number of DEGs (anderson-darling)']>50] |
| filtered = filtered[filtered['percent knockdown']<=-0.3] |
| filtered = filtered[filtered['number of cells (filtered)']>25] |
| strong_perts = filtered['genetic perturbation'].values |
| strong_perts = [s.split('_')[1] for s in strong_perts] |
| return strong_perts |
|
|
| def filter_cells_by_pert_effect(adata, k=10): |
|
|
| subset_idxs = [] |
| ctrl_adata = adata[adata.obs['gene'] == 'non-targeting'] |
|
|
| for itr, g in enumerate(adata.obs['gene'].unique()): |
| subset = adata[adata.obs['gene'] == g] |
|
|
| if g == 'non-targeting': |
| subset_idxs.append(subset.obs.index.values) |
| continue |
|
|
| try: |
| gene_loc = np.where(adata.var.gene_name == g)[0][0] |
| thresh = np.percentile(ctrl_adata.X[:,gene_loc],k) |
|
|
| subset_idxs.append(subset.obs.index[subset.X[:,gene_loc]<=thresh].values) |
| except: |
| subset_idxs.append(subset.obs.index.values) |
|
|
| subset_idxs = [item for sublist in subset_idxs for item in sublist] |
| filtered_adata = adata[subset_idxs,:] |
|
|
| return filtered_adata |
|
|
|
|
| def preprocess_dataset(dataset_path: str, summary_stats: pd.DataFrame = None): |
| """Preprocess the Anndata dataset and extract the necessary information |
| |
| Args: |
| dataset_path (string): path to Anndata dataset |
| summary_stats (pandas.DataFrame): dataframe containing summary stats to filter for strong perturbations. If None, do not filter. |
| |
| |
| Returns: |
| (numpy, list, list): expression matrix, list of gene ids (columns), list of perturbed genes (rows) |
| """ |
| data_expr_raw = sc.read(dataset_path) |
| if summary_stats is not None: |
| |
| strong_perts = get_strong_perts(summary_stats) + ['non-targeting'] |
| idx_to_keep = [v in strong_perts for v in data_expr_raw.obs['gene']] |
| pert_filter_k562 = data_expr_raw[idx_to_keep] |
| data_expr_raw = filter_cells_by_pert_effect(pert_filter_k562) |
|
|
| |
| sc.pp.normalize_per_cell(data_expr_raw, key_n_counts='UMI_count') |
| sc.pp.log1p(data_expr_raw) |
| data_expr_raw_df = data_expr_raw.to_df() |
| intervened_genes = list(data_expr_raw.obs["gene_id"]) |
| gene_to_interventions = dict() |
| for i, intervention in enumerate(intervened_genes): |
| if intervention in set(data_expr_raw_df.columns.to_list()) or intervention == "non-targeting": |
| gene_to_interventions.setdefault(intervention, []).append(i) |
| intervened_genes_set = set() |
| for gene, cell_indices in gene_to_interventions.items(): |
| if len(cell_indices) > 100: |
| intervened_genes_set.add(gene) |
| intervened_genes = ["excluded" if gene not in intervened_genes_set else gene for gene in intervened_genes] |
| data_expr_raw_perturbed_only = data_expr_raw_df[data_expr_raw_df.columns[data_expr_raw_df.columns.isin(intervened_genes_set)]] |
| gene_ids = data_expr_raw_perturbed_only.columns.to_list() |
| return data_expr_raw_perturbed_only.to_numpy(), gene_ids, intervened_genes |
|
|
