Request to add a GPC Phylogenetic Tree

[JoiRichi]

Context

The glycoprotein complex (GPC) is the only surface protein of the Lassa virus (LASV) and plays a crucial role in mediating entry into host cells. It is also a primary target for neutralizing antibodies and vaccine design efforts. Meanwhile, different lineages of LASV circulates in non-overlapping regions in West Africa and have differences in their GPC (Daodu et al. 2024). Understanding the phylogenetic relationships of GPC sequences can provide insights into the virus evolution, mechanisms of immune escape, and guide therapeutic development and distribution.

While segment-based trees (S and L) provide valuable information, a GPC-specific tree would complement these by offering a more detailed view of the evolutionary dynamics of the glycoprotein. This can be particularly useful in understanding the genetic diversity and evolutionary pressures acting on the GPC, which may not be fully captured by segment-based analyses.

In addition, a few recent papers have reported recombination within the virus segments, which may lead to inaccurate tree topologies when investigating gene specific questions (He et al. 2024; Wang et al. 2024).

Description

Currently, the Nextstrain build for LASV generates phylogenetic trees by segment (S or L). Is it possible to add generation of a phylogenetic tree based on the GPC of the Lassa virus?

Possible solution

• Using a GPC reference, manually created here: https://github.com/JoiRichi/LASV_phylogenetics_pipeline/blob/main/config/gly_ref_LASV.gb
• Create a GPC phylogenetic tree

[j23414]

From last week's discussion with @JoiRichi , there are some manual fixes necessary for the GPC alignment. Potential solutions include pulling the following manually fixed alignments from the following files as part of the GPC build pipeline:

1. Pull manually fixed alignment for GPC/sequences.fasta
2. Identify new sequences that are not in the manually curated alignment
3. Align new sequences using the augur align --existing-alignment feature to append new alignments
4. Thread through the rest of the phylogenetic pipeline as usual

[trvrb]

How many more sequences are available for glyoprotein than the entire S segment? GPC is ~1500 bases and the full S segment is ~3300 bases. Without more GPC-specific data you won't get a more detailed view of GPC evolution than what's visible in the S segment build https://nextstrain.org/staging/lassa/s. All amino acid changes to GPC are viewable, ie https://nextstrain.org/staging/lassa/s?c=gt-GPC_208. This is the same reason we don't need a spike-specific build for SARS-CoV-2.

[JoiRichi]

I believe there are more GPC sequences than full S segment sequences. Labs sometimes use sanger sequencing or partial sequencing of the gpc to track the virus. (For example, Adesina et al (2024) recently submitted over 80 GPC sequences. I doubt the corresponding NP was submitted as well.) Meanwhile, adding the nucleoprotein and the GPC together in a single S tree will definitely mask some of the information that can be gotten from standalone trees. Whitmer et al. (2018) already showed differences in tree topology across individual genes. Possible reasons: 1. Due to recombination (two studies published earlier this year showed evidence of intra-segmental recombination) 2. There may be differences in immune pressure acting on the different proteins because they respond to different antibodies (leading to variation in evolution). Having a standalone GPC tree would enable the community focus on the most important protein for immunotherapy and vaccines. Moreover, we have verified the accuracy of the alignment -which may be incorrect in some other LASV trees. For example, our latest preprint showed that over 90% of Nigerian LASV GPC have an indel in amino acid position 60 ( https://www.biorxiv.org/content/10.1101/2024.07.31.605963v1.abstract). This indel is missing in the current S tree ( https://nextstrain.org/staging/lassa/s?c=gt-GPC_60&gmax=1525&gmin=53) Overall, it may be important to have a verified GPC tree to assist the effort currently going in immunotherapy and vaccine developments in the field.

…

On Tue, 13 Aug 2024, 22:13 Trevor Bedford, ***@***.***> wrote: How many more sequences are available for glyoprotein than the entire S segment? GPC is ~1500 bases and the full S segment is ~3300 bases. Without more GPC-specific data you won't get a more detailed view of GPC evolution than what's visible in the S segment build https://nextstrain.org/staging/lassa/s. All amino acid changes to GPC are viewable, ie https://next.nextstrain.org/staging/lassa/s?c=gt-GPC_208 — Reply to this email directly, view it on GitHub <#20 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AOVPUJXVCXUM2SC2E2RKOI3ZRJZGJAVCNFSM6AAAAABME4QSD2VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDEOBXGE2DOMRZHA> . You are receiving this because you were mentioned.Message ID: ***@***.***>

[trvrb]

Thanks for the feedback @JoiRichi! This was very helpful to my understanding.

[trvrb]

there are some manual fixes necessary for the GPC alignment. Potential solutions include pulling the following manually fixed alignments from the following files as part of the GPC build pipeline:

This might be potentially addressed by codon-aware alignment via nextclade (or another codon-aware aligner that isn't MAFFT).