HLA, the coding genes of human major histocompatibility (MHC) proteins, play a crucial role in the human adaptive immune system by presenting antigenic peptides to T cell receptors on T cells. HLA-A, HLA-B and HLA-C, these 3 Class I HLA genes are one of the most polymorphic loci in the human genome. For decades, HLA typing has been performed prior to tissue and stem cell transplantation. However, beyond the role in tissue matching, HLA has also been implicated in a wide array of autoimmune diseases and HLA genotypes and expression levels are closely associated with cancer patients prognosis as recent studies have revealed. Recently methods have been developed to perform HLA typing and HLA expression quantification together by using RNA-seq techniques. However, these bulk RNA-seq experiments are measuring an averaged signal of cell populations. Single-cell RNA-seq (scRNA-seq) has regained its popularity due to its power to reliably resolve single RNA transcriptomes at large scales. In our present study, we did HLA typing using three independent scRNA-seq datasets. Interestingly, we found that single cells from the same donor could be classified into different groups where each group has a distinct expressed HLA genotype (e.g., HLA-A, heterozygous or homozygous); in other words, HLA class I genes show abundant allele specific expression in single cells. This phenomenon has been repeatedly observed in a total of 14 donors from 3 independent datasets (one is breast epithelium, another two are multiple myeloma). Our systematic analysis of HLA class I gene expression using multiple scRNA-seq datasets has uncovered a putative mechanism, where by fine tuning HLA class I expressions both at the quantity and allele levels, our immune system is able to handle various internal challenges through single cells equipped with extraordinary diverse HLA expression patterns.