Automatic detection of n-degree family members

Family-based genetic studies often require the identification of relatives up to a specified degree, but existing tools are either restricted to second-degree relatives, return entire connected pedigrees, or require multiple pre- or post-processing steps. We implemented five new functions, namely, <i>prepare_graph, get_kinship, graph_to_trio, get_relations</i>, and <i>Relation_per_proband_plot,</i> in the R package LTFHPlus to address these limitations. <i>prepare_graph</i> constructs a directed graph from population-level trio data using the <i>igraph</i> package and supports attaching additional attributes to individuals. From this graph, relatives of arbitrary degree can be identified efficiently. <i>get_kinship</i> calculates a kinship matrix for all individuals in a (sub)graph, and <i>graph_to_trio</i> reconstructs trio information from identified families, enabling downstream use with other pedigree tools. In addition, familial relations can be labelled from the graph using the function <i>get_relations</i>, and the total and average of each relation per proband can be plotted using <i>Relation_per_proband_plot</i>. Using the publicly available minnbreast dataset, we constructed a graph containing 28,081 individuals and 30,720 familial edges. Across 1,000 repetitions, the median run-time for identifying all relatives up to the third degree for 500 randomly selected individuals was 0.03 s, and kinship matrix calculation had a median run-time of 1.57 s (single-threaded execution). These functions provide a reproducible, scalable, and interoperable solution for integrating family information into genetic analyses.