A uniform survey of allele-specific binding and expression over 1000-Genomes-Project individuals

Jieming Chen, Joel Rozowsky, Timur R. Galeev, Arif Harmanci, Robert R. Kitchen, Jason Bedford, Alexej Abyzov, Yong Kong, Lynne Regan, Mark GersteinNature Communications 2016 · DOI

Abstract

Large-scale sequencing in the 1000 Genomes Project has revealed multitudes of single nucleotide variants (SNVs). Here, we provide insights into the functional effect of these variants using allele-specific behaviour. This can be assessed for an individual by mapping ChIP-seq and RNA-seq reads to a personal genome, and then measuring 'allelic imbalances' between the numbers of reads mapped to the paternal and maternal chromosomes. We annotate variants associated with allele-specific binding and expression in 382 individuals by uniformly processing 1,263 functional genomics data sets, developing approaches to reduce the heterogeneity between data sets due to overdispersion and mapping bias. Since many allelic variants are rare, aggregation across multiple individuals is necessary to identify broadly applicable 'allelic elements'. We also found SNVs for which we can anticipate allelic imbalance from the disruption of a binding motif. Our results serve as an allele-specific annotation for the 1000 Genomes variant catalogue and are distributed as an online resource (alleledb.gersteinlab.org).

Introduction

The 1000 Genomes Project provided a comprehensive catalog of human genetic variation, yet the functional implications of most identified variants remain elusive. A significant portion of these variants reside in non-coding regions, complicating their annotation. To address this, researchers have turned to allele-specific behavior as a proxy for functional impact. By comparing the maternal and paternal alleles within a single individual, one allele serves as a natural control for the other. Discrepancies in transcription factor binding (Allele-Specific Binding, ASB) or gene expression (Allele-Specific Expression, ASE) indicate that a variant likely has a functional consequence.

However, large-scale allele-specific analysis faces technical challenges, including reference bias—where mapping software favors the standard reference genome—and overdispersion, where data variability exceeds the expectations of standard binomial models.

Methodology: The AlleleDB Pipeline

To overcome these hurdles, a uniform processing pipeline was developed to analyze 1,263 functional genomics datasets (RNA-seq and ChIP-seq) from 382 individuals.

The uniform processing pipeline for detecting allele-specific behavior across 382 individuals.

The pipeline incorporates several key innovations:

  1. Personal Diploid Genomes: Instead of using a single reference, custom diploid genomes were constructed for each individual to eliminate reference bias.
  2. Beta-Binomial Modeling: To account for overdispersion, the team replaced standard binomial tests with a beta-binomial model. This adjustment significantly reduces false positives in noisy datasets by better fitting the empirical distribution of allelic counts.
  3. Ambiguous Mapping Bias (AMB) Filter: Simulations were used to identify and exclude reads that could map to multiple genomic locations, ensuring that observed imbalances were biological rather than artifacts of sequence homology.
  4. Data Aggregation: To handle rare variants, the study employed "collapsed" and "expanded" approaches to pool data across individuals, identifying "allelic elements"—genomic regions consistently showing imbalance.
Comparison of empirical distributions with binomial and beta-binomial models under low (a) and intermediate (b) overdispersion.

Key Results and Insights

The study resulted in the creation of AlleleDB, a massive resource annotating over 69,000 functional variants.

  • The Catalog: The pipeline detected 63,541 unique ASE SNVs and 6,136 unique ASB SNVs. These variants were significantly enriched in regulatory regions such as promoters, enhancers, and 5' UTRs.
  • Heritability: Analysis of a family trio demonstrated high heritability of allele-specific behavior (r ≈ 0.78 to 0.84), suggesting these imbalances are robust biological signals.
  • Motif Disruption: The researchers successfully predicted ASB by identifying variants that disrupt known transcription factor binding motifs. Variants that weakened a binding site consistently showed a corresponding decrease in binding for that allele.
  • Evolutionary Pressure: The data suggests that ASE variants may be under weaker purifying selection compared to non-ASE variants, while ASB variants exhibit distinct enrichment patterns, indicating they are functionally unique.

Discussion and Future Directions

While AlleleDB provides a powerful annotation tool, several limitations remain. The identified SNVs are associated with imbalances but are not necessarily the causal variants; they may be in linkage disequilibrium with the actual functional change. Furthermore, the reliance on lymphoblastoid cell lines means that many tissue-specific allelic effects may not be captured. Future work utilizing long-read sequencing is expected to further refine mapping and phasing, providing even greater resolution for allele-specific studies.

The complete resource is available at alleledb.gersteinlab.org.

Summary updated

Abstract

Large-scale sequencing in the 1000 Genomes Project has revealed multitudes of single nucleotide variants (SNVs). Here, we provide insights into the functional effect of these variants using allele-specific behaviour. This can be assessed for an individual by mapping ChIP-seq and RNA-seq reads to a personal genome, and then measuring 'allelic imbalances' between the numbers of reads mapped to the paternal and maternal chromosomes. We annotate variants associated with allele-specific binding and expression in 382 individuals by uniformly processing 1,263 functional genomics data sets, developing approaches to reduce the heterogeneity between data sets due to overdispersion and mapping bias. Since many allelic variants are rare, aggregation across multiple individuals is necessary to identify broadly applicable 'allelic elements'. We also found SNVs for which we can anticipate allelic imbalance from the disruption of a binding motif. Our results serve as an allele-specific annotation for the 1000 Genomes variant catalogue and are distributed as an online resource (alleledb.gersteinlab.org).
BibTeX
@inproceedings{chen2016uniform,
  author = {Jieming Chen and Joel Rozowsky and Timur R. Galeev and Arif Harmanci and Robert R. Kitchen and Jason Bedford and Alexej Abyzov and Yong Kong and Lynne Regan and Mark Gerstein},
  title = {{A uniform survey of allele-specific binding and expression over 1000-Genomes-Project individuals}},
  booktitle = {Nature Communications},
  year = {2016},
  doi = {10.1038/ncomms11101},
}
Figure