Analysis of protein-coding genetic variation in 60,706 humans

Monkol Lek, Konrad J. Karczewski, Eric Vallabh Minikel, Kaitlin E. Samocha, Eric Banks, Timothy R. Fennell, Anne O’Donnell‐Luria, James S. Ware, Andrew Hill, Beryl B. Cummings, Taru Tukiainen, Daniel P. Birnbaum, Jack A. Kosmicki, Laramie E. Duncan, Karol Estrada, Fengmei Zhao, James Zou, Emma Pierce‐Hoffman, Joanne Berghout, D.n. Cooper, Nicole Deflaux, Mark A. DePristo, Ron Do, Jason Flannick, Menachem Fromer, Laura D. Gauthier, Jackie Goldstein, Namrata Gupta, Daniel P. Howrigan, Adam Kieżun, Mitja Kurki, Ami Levy Moonshine, Pradeep Natarajan, Lorena Orozco, Gina M. Peloso, Ryan Poplin, Manuel A. Rivas, Valentín Ruano-Rubio, Samuel A. Rose, Douglas M. Ruderfer, Khalid Shakir, Peter D. Stenson, Christine Stevens, Brett Thomas, Grace Tiao, Maria T. Tusie-Luna, Ben Weisburd, Hong‐Hee Won, Dongmei Yu, David Altshuler, Diego Ardissino, Michael Boehnke, John Danesh, Stacey Donnelly, Roberto Elosúa, José C. Florez, Stacey Gabriel, Gad Getz, Stephen J. Glatt, Christina M. Hultman, Sekar Kathiresan, Markku Laakso, Steven A. McCarroll, Mark I. McCarthy, Dermot McGovern, Ruth McPherson, Benjamin M. Neale, Aarno Palotie, Shaun Purcell, Danish Saleheen, Jeremiah M. Scharf, Pamela Sklar, Patrick F. Sullivan, Jaakko Tuomilehto, Ming T. Tsuang, Hugh Watkins, James G. Wilson, Mark J. Daly, Daniel G. MacArthur

Abstract

Large-scale reference data sets of human genetic variation are critical for the medical and functional interpretation of DNA sequence changes. Here we describe the aggregation and analysis of high-quality exome (protein-coding region) DNA sequence data for 60,706 individuals of diverse ancestries generated as part of the Exome Aggregation Consortium (ExAC). This catalogue of human genetic diversity contains an average of one variant every eight bases of the exome, and provides direct evidence for the presence of widespread mutational recurrence. We have used this catalogue to calculate objective metrics of pathogenicity for sequence variants, and to identify genes subject to strong selection against various classes of mutation; identifying 3,230 genes with near-complete depletion of predicted protein-truncating variants, with 72% of these genes having no currently established human disease phenotype. Finally, we demonstrate that these data can be used for the efficient filtering of candidate disease-causing variants, and for the discovery of human 'knockout' variants in protein-coding genes.

Introduction

The 2016 landmark paper from the Exome Aggregation Consortium (ExAC) represents a massive leap in our understanding of human genetic variation. By aggregating and jointly analyzing the protein-coding regions (exomes) of over 60,000 individuals, the authors created a high-resolution map of human diversity that fundamentally changed how we interpret rare genetic diseases.

The Challenge: Interpreting Rare Variation

Interpreting a patient's genetic data is a "needle in a haystack" problem. Every human carries thousands of rare variants; the challenge for clinicians is determining which one causes a specific disease. Before ExAC, reference databases like the 1000 Genomes Project were too small to capture the full spectrum of rare variation. If a variant wasn't in those databases, it was often assumed to be "pathogenic." However, many variants are rare but harmless. Furthermore, we lacked a systematic way to know which genes are "essential"—meaning they are under such strong evolutionary pressure that any major mutation would be lethal or prevent reproduction.

Methods and Scale

The authors developed a unified pipeline to process over 91,000 exomes, eventually filtering down to a high-quality set of 60,706 individuals. Key methodological innovations include joint variant calling, the development of the pLI Score (Probability of being LoF Intolerant) to identify genes under strong selection, and the MAPS (Mutability-Adjusted Proportion of Singletons) metric to account for varying mutation rates across the genome.

Figure 1: Overview of the ExAC dataset, including cohort size by ancestry, variant frequency distribution, and evidence for mutational recurrence.

Key Insights

Mutational Recurrence

At this scale, the researchers observed "saturation." They found that for highly mutable sites (like CpG transitions), they had observed nearly every possible mutation that isn't lethal. This provides direct evidence of widespread mutational recurrence in the human population.

Figure 2: Proportion of possible mutations observed in ExAC, highlighting the high saturation of CpG transitions compared to other mutational contexts.

The "Essential" Gene List

They identified 3,230 genes that are highly intolerant to protein-truncating mutations. Surprisingly, 72% of these genes have no known human disease association, suggesting they represent a "dark matter" of essential biology yet to be explored.

Correcting the Literature

The study revealed that many variants previously reported as "pathogenic" are actually quite common in the general population. This allowed for the reclassification of hundreds of variants as likely benign, preventing potential misdiagnoses.

Key Results

  • High Density of Variation: One variant for every 8 base pairs of the exome.
  • Filtering Power: Using ExAC as a reference reduced the number of candidate "disease-causing" variants in a typical patient by sevenfold compared to previous databases.
  • Human "Knockouts": Identification of 1,550 genes where at least one individual was a "natural knockout."
  • Ancestry Resolution: Successful categorization of individuals into five major continental ancestries using Principal Component Analysis (PCA).

Limitations and Future Directions

While ExAC is a massive step forward, it has limitations:

  • Ascertainment Bias: Samples came from disease-focused studies.
  • Ancestry Gaps: Under-representation of Middle Eastern, Central Asian, and African populations.
  • Exome-Only: Focuses only on the 2% of the genome that codes for proteins.
  • Lack of Phenotypes: Limited clinical data for the individuals in the cohort.

The authors conclude that even larger datasets—reaching into the hundreds of thousands or millions—will be needed to fully characterize the rarest and most medically relevant parts of our genetic code.

Summary updated

Abstract

Large-scale reference data sets of human genetic variation are critical for the medical and functional interpretation of DNA sequence changes. Here we describe the aggregation and analysis of high-quality exome (protein-coding region) DNA sequence data for 60,706 individuals of diverse ancestries generated as part of the Exome Aggregation Consortium (ExAC). This catalogue of human genetic diversity contains an average of one variant every eight bases of the exome, and provides direct evidence for the presence of widespread mutational recurrence. We have used this catalogue to calculate objective metrics of pathogenicity for sequence variants, and to identify genes subject to strong selection against various classes of mutation; identifying 3,230 genes with near-complete depletion of predicted protein-truncating variants, with 72% of these genes having no currently established human disease phenotype. Finally, we demonstrate that these data can be used for the efficient filtering of candidate disease-causing variants, and for the discovery of human 'knockout' variants in protein-coding genes.
BibTeX
@inproceedings{lek2016analysis,
  author = {Monkol Lek and Konrad J. Karczewski and Eric Vallabh Minikel and Kaitlin E. Samocha and Eric Banks and Timothy R. Fennell and Anne O’Donnell‐Luria and James S. Ware and Andrew Hill and Beryl B. Cummings and Taru Tukiainen and Daniel P. Birnbaum and Jack A. Kosmicki and Laramie E. Duncan and Karol Estrada and Fengmei Zhao and James Zou and Emma Pierce‐Hoffman and Joanne Berghout and D.N. Cooper and Nicole Deflaux and Mark A. DePristo and Ron Do and Jason Flannick and Menachem Fromer and Laura D. Gauthier and Jackie Goldstein and Namrata Gupta and Daniel P. Howrigan and Adam Kieżun and Mitja Kurki and Ami Levy Moonshine and Pradeep Natarajan and Lorena Orozco and Gina M. Peloso and Ryan Poplin and Manuel A. Rivas and Valentín Ruano-Rubio and Samuel A. Rose and Douglas M. Ruderfer and Khalid Shakir and Peter D. Stenson and Christine Stevens and Brett Thomas and Grace Tiao and Maria T. Tusie-Luna and Ben Weisburd and Hong‐Hee Won and Dongmei Yu and David Altshuler and Diego Ardissino and Michael Boehnke and John Danesh and Stacey Donnelly and Roberto Elosúa and José C. Florez and Stacey Gabriel and Gad Getz and Stephen J. Glatt and Christina M. Hultman and Sekar Kathiresan and Markku Laakso and Steven A. McCarroll and Mark I. McCarthy and Dermot McGovern and Ruth McPherson and Benjamin M. Neale and Aarno Palotie and Shaun Purcell and Danish Saleheen and Jeremiah M. Scharf and Pamela Sklar and Patrick F. Sullivan and Jaakko Tuomilehto and Ming T. Tsuang and Hugh Watkins and James G. Wilson and Mark J. Daly and Daniel G. MacArthur},
  title = {{Analysis of protein-coding genetic variation in 60,706 humans}},
  booktitle = {Nature},
  year = {2016},
  doi = {10.1038/nature19057},
}
Figure