hla_typing: Align reads from (sc)RNAseq to known HLA reference alleles,...

In Close-your-eyes/scexpr: Lots of helpers and wrappers to analyze scRNAseq data sets

Align reads from (sc)RNAseq to known HLA reference alleles, count the number of matches and infer the HLA type

Description

This function tries to infer the HLA type by plotting the number of matching reads. Inspection of output-plots is seen as an alternative to statistical analysis. Please provide data (read and hla_ref) for only one gene at a time, e.g. only HLA-A or HLA-B or HLA-C. Run the function sequentially for every single gene. Not every allele-combination is easily inferred. Results may vary with respect to persuasiveness. It is assumed that inferring the p-group is sufficient. Also deeper typing becomes uncertain. The results of pairwise matches are hence plotted by p-groups.

Usage

hla_typing(
  hla_ref,
  reads,
  allele_diff = 5,
  top_n_pairwise_results = 50,
  hla_seq_colName = "seq_Exon2_3",
  read_seq_colName = "seq",
  hla_allele_colName = "allele",
  read_name_colName = "readName",
  p_group_colName = "p_group",
  g_group_colName = "g_group",
  lapply_fun = lapply,
  ...
)

Arguments

hla_ref	a data frame preferentially prepared with scexpr::hla_df_from_xml
reads	a data frame preferentially prepared with scexpr::reads_from_bam
allele_diff	maximum allowed difference (factor) of read abundances between two alleles; e.g. if highest abundant HLA-A allele 1 has 20 matched reads and diff_allele_hits_single_results = 5, then other alleles need to have at least 4 (=20/5) matched reads to be considered in the subsequent pairwise matching. In other words, what is the maximum expected/allowed expression difference of alleles HLA-A from father and HLA-A from mother. Intended to speed up the subsequent pairwise matching.
top_n_pairwise_results	top ranks of unique_explained_reads_rank and total_explained_reads_rank of pairwise matches used for plotting
hla_seq_colName	column name of hla sequences in hla_ref
read_seq_colName	column name read sequences in reads
hla_allele_colName	column name of allele names in hla_ref
read_name_colName	column name of read names in reads
p_group_colName	column name of p_group in hla_ref
g_group_colName	column name of g_group in hla_ref
lapply_fun	function name without quotes; lapply, pbapply::pblapply or parallel::mclapply are suggested
...	additional argument to the lapply function; mainly mc.cores when parallel::mclapply is chosen

Details

Reads are checked for perfect matches (hits) in all provided hla reference alleles (or a sub-sequence of them, e.g. only exon 2 and 3 which are subject to highest variation). For every hla allele the number of hits are counted - the allele can 'explain' a number of reads. This is done for every allele on its own. Then pairwise combinations of alleles are checked for the number of reads that they explain redundantly (double_explained_reads) or uniquely (uniquely_explained_reads) and in total (total_explained_reads). The allele-combinations with the highest ranks (combination of total_explained_reads and uniquely_explained_reads) likely reflect the cells' HLA type.

Value

list of data frames and ggplot2 objects which, upon visual inspection, may allow to infer hla type

Close-your-eyes/scexpr documentation built on April 21, 2023, 10:27 a.m.