While handling some multiplexed 10x data this past week, I realized that it had been a little while since I’d had to get per-barcode clonotype information to easily add into a Seurat object or other data format. I’ve also been using Haynes Heaton’s souporcell tool and have been very happy with the results, so in this example we’ll take some genotype information and concatenate that with our clonotype information. If you haven’t read the souporcell preprint, you should do so! The souporcell GitHub repo can be found here. Anyways, here’s a few lines that will get you where you need to go using data.table to read things in and tidyverse to quickly shape things:
# Import your libraries
library(data.table)
library(tidyverse)
# Import your clonotype data
clonotype <- fread('~/projects/filtered_contig_annotations.csv')
# Import your genotype data
genotype <- fread('~/projects/clusters.tsv')
# Filter your clonotype data
clonotype.filter <- clonotype %>% filter(productive==TRUE) # Keep productive rearrangements
clonotype.filter.again <- clonotype.filter %>% group_by(barcode) %>% mutate(clonotype=paste(cdr3,collapse=',')) %>% select(clonotype) %>% unique() # Collapse CDR3s across chains, keep unique clonotypes, and collapse by barcode
# Neatly table everything
my.clonotypes <- merge(genotype, clonotype.filter.again, by='barcode', all.x = TRUE)
# If you had 3 genotypes mixed in the same lane, here's a quick way to view some of your top clonotypes per genotype
my.clonotypes %>% filter(assignment==0) %>% select(clonotype) %>% table() %>% sort(.,decreasing = TRUE) %>% head()
my.clonotypes %>% filter(assignment==1) %>% select(clonotype) %>% table() %>% sort(.,decreasing = TRUE) %>% head()
my.clonotypes %>% filter(assignment==2) %>% select(clonotype) %>% table() %>% sort(.,decreasing = TRUE) %>% head()
