Merge pull request #20 from Tobi1kenobi/main

Small celltypist changes

README.md

@@ -4,8 +4,7 @@
 
 ## Introduction
 
-**nf-core/yascp** is a bioinformatics best-practice analysis pipeline for deconvolution, qc, clustering, integration of a single cell datasets.
-This is a large scale single-cell pipeline developed initially for processing Cardinal project samples, however it is applicable to any other scRNA analysis. The pipeline has been inspired by [deconvolution](https://github.com/wtsi-hgi/nf_scrna_deconvolution.git ), [cellbender](https://github.com/wtsi-hgi/nf_cellbender ) and [qc](https://github.com/wtsi-hgi/nf_qc_cluster/tree/main ) pipelines initially developed in Anderson lab. 
+**nf-core/yascp** is a bioinformatics best-practice analysis pipeline tailored for deconvolution, quality control, clustering, and integration of single-cell datasets. Developed under the leadership of N.Soranzo and Human Genetics Informatics (HGI), this large-scale single-cell pipeline was originally crafted for the Cardinal project (profiling UKBB and ELGH participants) but is versatile enough for broader scRNA analysis applications. The foundational ideas were inspired by earlier pipelines from Anderson lab but has been expanded, specifically those for [deconvolution](https://github.com/wtsi-hgi/nf_scrna_deconvolution.git), [cellbender](https://github.com/wtsi-hgi/nf_cellbender), and [quality control and clustering](https://github.com/wtsi-hgi/nf_qc_cluster/tree/main). This ensures a robust integration of proven methodologies tailored to meet the demands of expansive single-cell data analysis.
 
 Input requires a tsv seperated file [(please read detailed documentation here)](https://github.com/wtsi-hgi/yascp/tree/yascp_docs) with paths and if running in an genotype  additional input is required to be provided in an input.nf file pointing to the vcf location. This pipeline is designed to be used any large scale single cell experiments.
 

bin/run_celltypist.py

@@ -145,7 +145,7 @@ def run_celltypist(samplename, filtered_matrix_h5, celltypist_model,
 
     # Indeed, the `model` argument defaults to `Immune_All_Low.pkl`.
     logging.info("celltypist_model: " + celltypist_model)
-    celltypist_model1 = celltypist_model.split('.')[0]
+    celltypist_model1 = os.path.splitext(os.path.basename(celltypist_model))[0]
     # print(os.listdir('/tmp/.celltypist/data/models'))
     model = models.Model.load(model = celltypist_model ) # model = 'Immune_All_Low.pkl')
     model.description
@@ -172,7 +172,6 @@ def run_celltypist(samplename, filtered_matrix_h5, celltypist_model,
 
     adata = predictions.to_adata()
 
-
     predictions.to_table(folder = output_dir, prefix = samplename + '___'+celltypist_model1+'___')
     Data = adata.obs
     Data= Data.drop('cell_barcode',axis=1)

modules/nf-core/modules/celltypist/main.nf

@@ -1,6 +1,6 @@
 
 process CELLTYPIST {
-    tag "${samplename}"
+    tag "${model}_${sample}"
     label 'process_medium_memory'
     publishDir "${params.outdir}/celltype/celltypist/${model}/${sample}/", mode: "${params.celltypist.copy_mode}", overwrite: true,
 	  saveAs: {filename -> filename.replaceFirst("outputs/","").replaceFirst("figures/","") }
@@ -27,7 +27,7 @@ process CELLTYPIST {
       tuple val(sample), path("outputs/plot_prob/*_*.pdf"), emit: sample_plots_prob_pdf
 
     script:
-      model="${celltypist_model}".replaceFirst(".pkl","")
+      model="${celltypist_model}".replaceAll(/^.*[\\/]/, "").replaceFirst(".pkl","")
 
       filtered_matrix_h5_path = file("${filtered_matrix_h5}/../filtered_feature_bc_matrix.h5")
       if (filtered_matrix_h5_path.exists()){