HIV molecular immunology database

This tool takes a series of tab-delimited HLA types and returns a graph representing this data. For more information, see details below, and try Sample Input.

Input

This program expects a list of HLA Class-I alleles. Each line of the input should be a unique patient ID followed by a list of HLAs, all separated by spaces or tabs. Missing data may be omitted, or indicated by a dash '-'.

HLA frequencies can be counted two different ways, which differ in how homozygous alleles are counted.

- Population:
- Frequencies of HLA type per person. Homozygous HLAs are counted once.
- Allele:
- Frequencies of HLA type per allele. Homozygous HLAs are counted twice.

HLA types are reported with various resolutions (2, 4 or more digits). This tool provides three options for handling this.

- Leave HLAs as presented:
- The HLA type is left unchanged and each distinct type is counted separately.
- 4-digit HLA comparison:
- Silent mutations are ignored; HLA types with more than 4 digits precision are truncated. HLA types with 2 digits are left unchanged.
- 2-digit HLA comparison:
- All HLA types are converted to 2 digits. The conversion is based on serotype information, if available, or else on the 2-digit HLA family.

In the results, the contingency table (column C) reports the
number of subjects in each population with and without the given
HLA. For each HLA, the tool computes the 2-sided exact Fisher's
p-value, which represents the probability that the observed
difference is due to chance. To correct for the false discovery
rate caused by the calculation of multiple p-values, Storey's
q-value is also provided (Storey JD: A direct approach to false
discovery rates. *J R Statist Soc B* 2002, 64:479-498).