Code for gnomAD LoF flagship manuscript

Manuscript at: Karczewski et al., 2019

Details

This repo serves as a home for two main purposes: - The constraint computation pipeline, written in Hail 0.2. - The figure-generating code for the manuscript

Constraint

The constraint pipeline, as initially described here and here, has been updated with a number of improvements as described in the supplement of Karczewski et al. Notably, the pipeline is now written in Hail, which enables scalability to large datasets like gnomAD, and can compute constraint against arbitrary sets of variants.

The main components of the pipeline can be found in constraint/constraint.py, which uses the public gnomAD data and a dataset of every possible variant (~9B variants) to compute the observed and expected number of variants per transcript/gene. This script is provided primarily for reference (it could be run with modifications, but cannot be run as-is, as it has paths to buckets on Google cloud hard-coded). Additionally, we combine the constraint data with aggregate LoF frequencies in constraint/gene_lof_matrix.py - this script cannot be run outside of the gnomAD team, as it requires access to the individual level data, but it is also provided for reference.

Figure-generating code

The code to generate all the figures can be found in R/. These scripts use only aggregated data, and thus, can be run by anyone. They look for data files in ./data, and if they are not found, downloads them as needed from the public data repository.

Each figure panel can be generated individually, or figures as a whole. For instance, in R/fig3_spectrum.R, we provide a function called figure3 which can generate the entirety of figure 3. Alternatively, running the code inside the function can generate each figure panel separately. Note that for some figures, on some R setups, attempting to generate the full figure by calling the function directly can crash R: we are uncertain of the cause of the issue, but it can be resolved by running the code inside the function step-wise.

The code was run using R 3.5.1 using the following packages:

```

sessionInfo() R version 3.5.1 (2018-07-02) Platform: x86_64-apple-darwin18.2.0 (64-bit) Running under: macOS 10.14.3

Matrix products: default BLAS: /System/Library/Frameworks/Accelerate.framework/Versions/A/Frameworks/vecLib.framework/Versions/A/libBLAS.dylib LAPACK: /opt/local/Library/Frameworks/R.framework/Versions/3.5/Resources/lib/libRlapack.dylib

locale: [1] enUS.UTF-8/enUS.UTF-8/enUS.UTF-8/C/enUS.UTF-8/en_US.UTF-8

attached base packages: [1] grid stats graphics grDevices utils datasets methods base

other attached packages: [1] hexbin1.27.2 bindrcpp0.2.2 cowplot0.9.4 RMySQL0.10.16 DBI1.0.0
[6] ggrepel0.8.0 pbapply1.3-4 rlang0.3.1 tidygraph1.1.1 STRINGdb1.22.0
[11] meta4.9-4 ggrastr0.1.7 ggpubr0.2 ggridges0.5.1 readxl1.2.0
[16] corrr0.3.0 corrplot0.84 patchwork0.0.1 naniar0.4.1 plotROC2.2.1
[21] gghighlight0.1.0 skimr1.0.4 gapminder0.3.0 trelliscopejs0.1.18 scales1.0.0
[26] magrittr1.5 slackr1.4.2 plotly4.8.0 broom0.5.1 forcats0.3.0
[31] stringr1.3.1 dplyr0.7.8 purrr0.2.5 readr1.3.1 tidyr0.8.2
[36] tibble2.0.1 tidyverse1.2.1 Hmisc4.1-1 ggplot23.1.0 Formula1.2-3
[41] survival2.43-3 lattice0.20-38

loaded via a namespace (and not attached): [1] colorspace1.4-0 visdat0.5.2 mclust5.4.2 htmlTable1.13.1
[5] base64enc0.1-3 rstudioapi0.9.0 hash2.2.6 bit640.9-7
[9] fansi0.4.0 lubridate1.7.4 sqldf0.4-11 xml21.2.0
[13] codetools0.2-16 splines3.5.1 knitr1.21 jsonlite1.6
[17] Cairo1.5-9 cluster2.0.7-1 png0.1-7 compiler3.5.1
[21] httr1.4.0 backports1.1.3 assertthat0.2.0 Matrix1.2-15
[25] lazyeval0.2.1 cli1.0.1 acepack1.4.1 htmltools0.3.6
[29] prettyunits1.0.2 tools3.5.1 igraph1.2.2 gtable0.2.0
[33] glue1.3.0 Rcpp1.0.0 cellranger1.1.0 gdata2.18.0
[37] nlme3.1-137 autocogs0.1.1 xfun0.4 proto1.0.0
[41] rvest0.3.2 gtools3.8.1 DistributionUtils0.6-0 hms0.4.2
[45] parallel3.5.1 RColorBrewer1.1-2 yaml2.2.0 memoise1.1.0
[49] gridExtra2.3 rpart4.1-13 latticeExtra0.6-28 stringi1.2.4
[53] RSQLite2.1.1 plotrix3.7-4 checkmate1.9.1 caTools1.17.1.1
[57] chron2.3-53 pkgconfig2.0.2 bitops1.0-6 bindr0.1.1
[61] labeling0.3 htmlwidgets1.3 bit1.1-14 tidyselect0.2.5
[65] plyr1.8.4 R62.3.0 gplots3.0.1 generics0.0.2
[69] gsubfn0.7 pillar1.3.1 haven2.0.0 foreign0.8-71
[73] withr2.1.2 RCurl1.95-4.11 nnet7.3-12 modelr0.1.2
[77] crayon1.3.4 utf81.1.4 KernSmooth2.23-15 progress1.2.0
[81] data.table1.12.0 blob1.1.1 digest0.6.18 webshot0.5.1
[85] munsell0.5.0 viridisLite0.3.0 egg_0.4.2
```