Estimating the potential to prevent locally acquired HIV infections in a UNAIDS Fast-Track City, Amsterdam

This repository includes code and partial data for the analyses in Blenkinsop, Monod, van Sighem et al. eLife (2022) https://doi.org/10.7554/eLife.76487

Data

The data folder contains subfolders with the following input files:

• trees
  • reconstructed phylogenetic trees labelled with risk group and year of sequence sample
• subgraphs
  • Amsterdam subgraphs extracted from trees
• subgraph_metatadata
  • classification of subgraphs as pre-existing by 2014 or emergent since 2014
• patient data
  • file containing a flag to indicate whether patient ID was virally suppressed by 2014
  • file containing a flag to indicate whether a patient had an estimated infection date after 2014
  • number of diagnosed individuals estimated to have been infected since 2014 by transmission risk group and place of birth
  • number of sequenced individuals estimated to have been infected since 2014 by transmission risk group, place of birth and HIV subtype
• infection times
  • estimated time-to-diagnosis data by risk group and migrant group
  • estimated infected individuals by year for MSM/non-MSM in Amsterdam from the European Centres for Disease Control (ECDC) HIV modelling tool

Code

The analysis is run in 3 stages. The first stage requires sequence data and patient meta-data. The second two stages can be run using aggregated patient and phylogenetic data in the repository to replicate the results of the manuscript.

Phylogenetic analysis

1. scripts/pre-process-sequences.R - Pre-processes sequence data
2. scripts/phylo-analysis.R - Runs phylogenetic analysis

Estimating the undiagnosed population

1. run-undiagnosed.R - Writes a shell script to estimate the proportion undiagnosed.

Estimating locally acquired infections

1. submit-job-MSM.R & submit-job-HSX.R - Writes shell script to run the analysis on a computer cluster including preparing the stan data (stan-make-data.R), sampling using cmdstan, and post-processing. Requires the job name of the undiagnosed model as input.
2. scripts/post-processing-combine-MSM-HSX-results.R - Writes a shell script to combine the results of the independent MSM and heterosexual model. Both jobs must have completed before running.