The HLA Imputation Portal (HIP) consumes .bed, .bim, and .fam files for SNP-genotyped subjects, and imputes HLA-A, -C, -B, -DRB1, -DQA1, -DQB1, and -DPB1 genotypes for each subject. The HIP does not consume .ped and .map files. PLINK can be used to convert .ped and .map files to .bed, .bim and .fam files on your system (http://zzz.bwh.harvard.edu/plink/data.shtml#bed).

The HIP can only be applied to subjects with origin identifiers that have previously been loaded into the HLA|COVID-19 Database (HCDB) under a specific Project Name. To create a HCDB account and/or load origin identifiers and clinical data for subjects, visit https://database-hlacovid19.org.

SNPs must be described using hg19 SNP rs IDs and genomic coordinates.

The HIP relies on 129 models built from genome-wide SNP data generated on 30 array platforms for subsets of African, Asian, European, Hispanic and Multi-ethnic groups. For details of the platforms and the associated population models, visit https://zhengxwen.github.io/HIBAG/platforms.html.

To access the HIP, go to https://database-hlacovid19.org/shiny in your browser, and click the “HLA-Imputation-Portal/” link. Once connected to the HIP, provide your HCDB account’s associated email address, and the Project Name associated with the origin identifiers for each subject in your SNP dataset(shown below). As noted above, these origin identifiers must be included in the uploaded .fam file, and must have been loaded into the HCDB before using the HIP.

Select the pertinent population group in the “Ethnic population” pulldown menu, and the pertinent SNP genotyping assay in the “Genotyping Method” pulldown menu. Note that not all combinations of ethnic population and genotyping assay are possible. If a specific population group is not available for your genotyping assay, choose the “Multi-ethnic/Other” group.

Click on the “Browse…” button for each of the .bed, .bim and .fam files to select the pertinent file for imputation. These data are extracted from your local files; thefiles are not uploaded to the HIP.

When the submitter’s email and the Project Name have been provided, the population and assay have been selected, and the source files have been successfully uploaded, press the “Impute!” button to begin the imputation process. Any messages about the imputation process will be displayed in the HIP’s Main tab. Specific information about these messages are provided in the About tab.

While the imputation is running, a circle of spinning dots will appear in the upper-right of the page, the HIP page will be greyed out, and a white bar with the message “Your imputation is running…” will appear near the top of the page.

When the imputation is complete, a message will appear in the Main tab, and an email notification will be sent to the address provided in the “Submitter’s e-mail” field. Imputed HLA genotypes for each subject will be automatically loaded into the specified Project Name of the HCDB account associated with the Submitter’s e-mail address for the origin identifiers in the .fam file. In addition, a downloadable file of posterior probability values for each subject and each locus will be loaded into your account.

Because the imputation process relies on specific ethnic models, it may be most efficient to generate separate .fam files for subjects of different ethnicities, and run each .fam file separately. Imputation performed using the Multi-ethnic models can take significantly longer than using the single-ethnicity models, and use of the African, Asian, European and Hispanic models when possible is recommended. In instances where a specific ethnic model is not available for a particular SNP assay, use the Multi-ethnic model.