HIV Databases HIV Databases home HIV Databases home
HIV sequence database

Glycan Shield Mapping Explanation

Input Data

This tool takes as input an amino acid sequence alignment of the HIV Env sequences for glycan shield prediction. If possible, please include HXB2 as the reference sequence in this alignment (it should be called exactly "HXB2"). While the tool does provide an option to add HXB2, user-added HXB2 is preferred so that the user can control how the Envs of interest are aligned to HXB2. To generate an alignment of your sequences, please use HIVAlign. HXB2 can also be added using this tool.

It is recommended that the user checks the resulting alignment manually to ensure proper alignment especially in the hypervariable regions. The preferred way of aligning hypervariable regions is to evenly split the hypervariable loops close to each end, e.g. for hypervariable V1 HXB2 positions 132-152

BG505:      TNVTNNI--------TDDMRG
The HXB2 positions of hypervariable loops are V1: 132-152, V2: 185-190, V4: 396-410, V5: 460-465 (V3 does not have a hypervariable region).

Run Time

Please limit the number of sequences given, each sequence takes approximately 5 mins. to analyze. For more than 2 sequences (HXB2 excluded) please use the 'Email results' option.


This tool has two options. The first option queries whether the user-provided input alignment contains the HXB2 reference sequence. If the user clicks "No" then this tool automatically adds and aligns HXB2. The second option is to send an email after the analyses are done (the link to the output will be provided in the email). This option is recommended when analyzing more than 2 sequences (excluding HXB2).

Output Page

The output page shows results for each sequence in the input alignment (reference sequence HXB2 is not analyzed). Next to the sequence name is the predicted total glycan hole area (using the >50% M-group conserved glycan shield as baseline, see Wagh et al., 2018 ; units Å2 per trimer). The first row of figures shows the predicted glycan shield from 3 points of view: a) "front" is the view with viral membrane at the bottom; b) "side" rotates the previous view ~120° degrees counterclockwise around the vertical axis; and c) "top" shows the view along the trimer axis from the trimer apex.

This tool also predicts the potential N-linked glycosylation sites (PNGS), which are commonly found in HIV-1 group M Envs but absent from the given Env, that gave rise to the glycan holes. The common M-group PNGS used are the HXB2 positions: 88, 130, 156, 160, 197, 234, 241, 262, 276, 289, 295, 301, 332, 339, 356, 386, 392, 448, 611, 616, 625, 637. The second row of figures shows the mapping of the absent hole-causing PNGS on the predicted glycan shields – the PNGS are shown in cyan with their position labeled in black on the structure, and their 10Å neighbors are shown in yellow. Please note, glycan shielding in hypervariable loops cannot be reliably predicted due to very high variation in sequence and length, and in structural orientations. While these regions are shown, glycan shield mapping in these regions should be treated with caution, and the PNGS that cause the holes in these regions are not analyzed.

The high-resolution figures can be downloaded by clicking the image, or by downloading all the results using a link at the top of the output webpage.

The figures are followed by a list of sites (using HXB2 numbering) that comprise all the predicted glycan holes in the given Env sequence. For this calculation, only those sites that have at least a single heavy atom exposed on the surface of the reference trimer structure are reported, and hypervariable sites are ignored.

Following this a table that shows data on the extent of holes due to each absent PNGS. The 2nd column shows the hole area accounted for by adding each missing PNGS to the given Env sequence; the absent PNGS are ordered using this metric with the biggest effect PNGS on the top. The 3rd column shows the hole area accounted for by a given PNGS in the presence of all the PNGS above it. If two absent PNGS are such that their predicted glycan shields overlap, then the total hole area accounted by the second PNGS will depend on whether the first PNGS is present or absent. For example, JRFL has missing PNGS at N289 and N234. The first alone accounts for 1641.65Å2, while the second alone accounts for 1145.59Å2; however, N234 in the presence of N289 accounts for 965.01Å2. The fourth column lists the sites in the glycan holes that could be covered by the given missing PNGS. Please note that a single site could be covered by two or more missing PNGS if they are structurally proximal.

Results zip archive

All the output figures, data and tables can be downloaded in a zip archive using the link provided at the top of the output page. The figures are named using the sequence name, view ("front", "side" or "top") and whether or not missing PNGS are labeled. The tables are provided in .csv format.

In addition, three additional files are provided. These are related to the Pymol scripting that is used to generate the structure images and include: a) the reference structure (PDB: 5FYJ, modified as in Wagh et al. ), b) the Pymol script ("pymol_script.pml") and c) output from Pymol ("pymol_out.txt"). If the user has Pymol installed, they can locally run the script to generate the outputs.

last modified: Wed Nov 21 00:21 2018

Questions or comments? Contact us at

Operated by Triad National Security, LLC for the U.S. Department of Energy's National Nuclear Security Administration
© Copyright Triad National Security, LLC. All Rights Reserved | Disclaimer/Privacy

Dept of Health & Human Services Los Alamos National Institutes of Health