A Comprehensive Toolkit for Clinical HLA Informatics.

immunogenetr

immunogenetr is a comprehensive toolkit for clinical HLA informatics. It is built on tidyverse principles and makes use of genotype list string (GL string, https://glstring.org/) for storing and using HLA genotype data.

Specific functionalities of this library include:

• Coercion of HLA data in tabular format to and from GL string.
• Calculation of matching and mismatching in all directions, with multiple output formats.
• Automatic formatting of HLA data for searching within a GL string.
• Truncation of molecular HLA data to a specific number of fields.
• Reading HLA genotypes in HML files and extracting the GL string.

Table of Contents

• Installation
• Usage
• License
• Disclaimer

Installation

You may install immunogenetr from CRAN with the below line of code:

install.packages("immunogenetr")

Usage

To demonstrate some functionality of immunogenetr we will use an internal dataset to perform match grades for a putative recipient/donor pair.

library(immunogenetr)
library(tidyverse)

# The "HLA_typing_1" dataset is installed with immunogenetr, and contains high resolution typing at all classical 
# HLA loci for ten individuals.

print(HLA_typing_1)

immunogenetr uses genotype list strings (GL strings) for most functions, including the matching and mismatching functions. To easily convert the genotypes found in “HLA_typing_1” to GL strings we can use the HLA_columns_to_GLstring function:

HLA_typing_1_GLstring <- HLA_typing_1 %>% 
  mutate(GL_string = HLA_columns_to_GLstring(., HLA_typing_columns = A1:DPB1_2), .after = patient) %>% 
  # Note the syntax for the `HLA_columns_to_GLstring` arguments - when this function is used inside 
  # of a `mutate` function to make a new column in a data frame, "." is used in the first argument 
  # to tell the function to use the working data frame as the source of the HLA typing columns.
  select(patient, GL_string)

print(HLA_typing_1_GLstring)

The “HLA_typing_1_GLstring” data frame now contains a row with a GL string for each individual, containing their full HLA genotype in a single string. Let’s select one individual to act as a recipient, and one to act as a donor.

# Select one case each for recipient and donor.
HLA_typing_1_GLstring_recipient <- HLA_typing_1_GLstring %>% 
  filter(patient == 7) %>% 
  rename(GL_string_recipient = GL_string, case = patient)

HLA_typing_1_GLstring_donor <- HLA_typing_1_GLstring %>% 
  filter(patient == 9) %>% 
  rename(GL_string_donor = GL_string) %>% 
  select(-patient)

# Combine the tables so recipient and donor are on the same row.
HLA_typing_1_recip_donor <- bind_cols(
  HLA_typing_1_GLstring_recipient, 
  HLA_typing_1_GLstring_donor
  )

print(HLA_typing_1_recip_donor)

We now have a data frame with a recipient and donor HLA genotype on one row. Let’s try out some of the mismatching functions on this data.

HLA_typing_1_recip_donor_mismatches <- HLA_typing_1_recip_donor %>% 
  mutate(A_MM_GvH = HLA_mismatch_logical(
                      GL_string_recipient, 
                      GL_string_donor, 
                      "HLA-A", 
                      direction = "GvH"), 
                    .after = case) %>% 
  mutate(A_MM_HvG = HLA_mismatch_logical(
                      GL_string_recipient, 
                      GL_string_donor, 
                      "HLA-A", 
                      direction = "HvG"), 
                    .after = A_MM_GvH)

print(HLA_typing_1_recip_donor_mismatches)

The HLA_mismatch_logical function determines if there are any mismatches at a particular locus. We’ve determined that at the HLA-A locus there are not any mismatches in the graft-versus-host direction, but are in the host-versus-graft direction. We can use the HLA_mismatched_alleles function to tell us what those mismatches are:

HLA_typing_1_recip_donor_mismatched_allles <- HLA_typing_1_recip_donor %>% 
  mutate(A_HvG_MMs = HLA_mismatched_alleles(
                        GL_string_recipient, 
                        GL_string_donor, 
                        "HLA-A", 
                        direction = "HvG"), 
                      .after = case)

print(HLA_typing_1_recip_donor_mismatched_allles)

The HLA_mismatched_alleles function reported that the “HLA-A*30:01” allele was mismatched in the HvG direction. Sometimes, however, we simply want to know how many mismatches are at a particular locus. We can do that with the HLA_mismatch_number function:

# Determine the number of bidirectional mismatches at several loci.
HLA_typing_1_recip_donor_MM_number <- HLA_typing_1_recip_donor %>% 
  mutate(ABCDRB1_MM = HLA_mismatch_number(
                        GL_string_recipient, 
                        GL_string_donor, 
                        c("HLA-A", "HLA-B", "HLA-C", "HLA-DRB1"), 
                        direction = "bidirectional"), 
                      .after = case)

print(HLA_typing_1_recip_donor_MM_number)

We might want to calculate an HLA match summary for stem cell transplantation. We can use the HLA_match_summarry_HCT function for this:

# The match_grade argument of "Xof8" will return the number of matches at the HLA-A, B, C, and DRB1 loci.
HLA_typing_1_recip_donor_8of8_matching <- HLA_typing_1_recip_donor %>% 
  mutate(ABCDRB1_matching = HLA_match_summary_HCT(
                              GL_string_recipient, 
                              GL_string_donor, 
                              direction = "bidirectional", 
                              match_grade = "Xof8"), 
                            .after = case)

print(HLA_typing_1_recip_donor_8of8_matching)

Clearly, this recipient and donor are not a great match. Let’s see how we could use this workflow to find the best-matched donor from several options. To do this, we’ll choose a case from “HLA_typing_1” and compare it to all the cases in that data set:

# Select one case to be the recipient.
HLA_typing_1_GLstring_candidate <- HLA_typing_1_GLstring %>% 
  filter(patient == 3) %>% 
  select(GL_string) %>% 
  rename(GL_string_recip = GL_string)

# Join the recipient to the 10-donor list and perform matching
HLA_typing_1_GLstring_donors <- HLA_typing_1_GLstring %>% 
  rename(GL_string_donor = GL_string, donor = patient) %>% 
  cross_join(HLA_typing_1_GLstring_candidate) %>%
  mutate(ABCDRB1_matching = HLA_match_summary_HCT(
                              GL_string_recip, 
                              GL_string_donor, 
                              direction = "bidirectional", 
                              match_grade = "Xof8"), 
                            .after = donor) %>%
  arrange(desc(ABCDRB1_matching))

print(HLA_typing_1_GLstring_donors)

We can see that donor 3 is the only donor with an 8/8 match for the recipient.

License

This project is licensed under the GNU General Public License v3.0.

Disclaimer

This library is intended for research use. Any application making use of this package in a clinical setting will need to be independently validated according to local regulations.