OBJECTIVETo investigate the role of different immunoglobulin- like receptor (KIR)haplotypes in haplo- identical hematopoietic stem cell transplantation (HSCT).

METHODKiller cell KIR genotyping was performed on 468 individuals from 156 unrelated families by PCR-SSP. A total of 624 KIR haplotypes from the parents were used for haplotype analysis. Ninety-two patients received haplo-identical HSCT from one of the parents.

RESULTSThe family study showed segregation of one A haplotype and at least 20 unique B haplotypes. The frequency of haplotype A was 72.92% (455/624). The most commonly observed haplotypes in group B were B1, B2, and B3, present at a frequency of 10.26%, 5.77%, and 4.48%, respectively. Compared to KIR gene matched donors (n=17), grafts from KIR gene mismatched donors (n= 14) had a positive effect on survival after haplo- identical HSCT for AML/MDS patients (OS: 88.2%vs 42.9%,P=0.015; RFS: 88.2%vs 35.7%,P=0.007). No effect was observed for ALL/NHL patients (OS: 76.0%vs 75.0%,P=0.727; RFS: 68.0%vs 65.0%,P=0.866). A significantly lower survival rate was observed for transplants from AA (n=52) and AB1/AB2 donors (n=15), compared to other group Bx donors (n=25) (OS: 53.3%vs 96.0%,P=0.017; RFS: 53.3%vs 92.0%,P=0.019). Meanwhile, the risk of relapse was much higher in AA group (n=52) compared to Bx group (n=40) (25.0%vs 5.0%,P=0.009). A higher risk of TRM was observed in AB1/AB2 group (P=0.012). In addition, transplant from donors carried Cen-B was associated with an increased survival compared with Cen-A homozygous donors (OS: 94.7%vs 68.5%,P=0.036; RFS: 89.5%vs 64.4%,P=0.045).

CONCLUSIONOverall, KIR genotyping and haplotype analyses should be useful for selection of the most optimal donors with favorable KIR gene grafts. KIR gene mismatch donors should be preferred for AML/MDS patients. Selecting donors carried Cen- B and avoiding the selection of donors of KIR genotype AA/AB1/AB2 was strongly advisable for haplo-identical HSCT.

Chronic Disease ; Genotype ; Haplotypes ; Hematopoietic Stem Cell Transplantation ; Humans ; Killer Cells, Natural ; Leukemia, Myeloid, Acute ; therapy ; Neoplasm Recurrence, Local ; Receptors, KIR ; genetics ; Survival Rate ; Tissue Donors

Objective: To analyze family-based haplotype frequencies of HLA-A, -B, -C, -DRB1 and -DQB1 genes and their clinical significance. Methods: The data of HLA genotyping in 3568 families undergoing related haploidentical transplantation between 2012 and 2017 at the First Affiliated Hospital of Soochow University were retrospectively evaluated. The HLA genotyping was performed by PCR amplification with sequence-based typing (PCR-SBT) and sequence-specific oligonucleotide probe (PCR-SSOP) methods. The family genetic analysis and haplotype frequencies were also investigated. Results: All the families were divided into 3 groups, including group1 of 1 422 entire families; group2 of 1 310 patients and either of their parents or one of their children; group3 of 836 patients and their HLA≥5/10 matched sibling donors. In the haplotypes with frequencies greater than 0.1% in group1+ group2, the frequency of A*11∶01-B*40∶01-C*03∶04-DRB1*11∶01-DQB1*03∶01, A*02∶07-B*51∶01-C*14∶02-DRB1*09:01-DQB1*03∶03 were significantly different between group1 and group2 (P=0.029, 0.033) . The frequency of A*11∶01-B*46∶01-C*01∶02∶01G-DRB1*09∶01-DQB1*03∶03 was significantly different between group1 and group3 (P=0.035) . The frequency of A*02∶01-B*40∶01-C*07∶02-DRB1*09∶01-DQB1*03∶03 was significantly different between group1 and group2 (P=0.034) , or group1 and group3 (P=0.034) . The frequency of A*24∶02-B*13∶01-C*03∶04-DRB1*12∶02-DQB1*03:01 was significantly different between group2 and group3 (P=0.046) . Conclusion In this study, we summarize the prevalence of haplotype frequencies in terms of HLA-A, -B, -C, -DRB1 and-DQB1. Based on the database of family haplotype analysis, patients and donor candidates are sorted with matched HLA genotype while unmatched HLA haplotype. Even in patients without entire family information, HLA haplotype analysis assists in choosing the optimal related or unrelated donors.

Prediction of the interactions between small molecules and their targets play important roles in various applications of drug development, such as lead discovery, drug repurposing and elucidation of potential drug side effects. Therefore, a variety of machine learning-based models have been developed to predict these interactions. In this study, a model called auxiliary multi-task graph isomorphism network with uncertainty weighting (AMGU) was developed to predict the inhibitory activities of small molecules against 204 different kinases based on the multi-task Graph Isomorphism Network (MT-GIN) with the auxiliary learning and uncertainty weighting strategy. The calculation results illustrate that the AMGU model outperformed the descriptor-based models and state-of-the-art graph neural networks (GNN) models on the internal test set. Furthermore, it also exhibited much better performance on two external test sets, suggesting that the AMGU model has enhanced generalizability due to its great transfer learning capacity. Then, a naïve model-agnostic interpretable method for GNN called edges masking was devised to explain the underlying predictive mechanisms, and the consistency of the interpretability results for 5 typical epidermal growth factor receptor (EGFR) inhibitors with their structure‒activity relationships could be observed. Finally, a free online web server called KIP was developed to predict the kinome-wide polypharmacology effects of small molecules (http://cadd.zju.edu.cn/kip).