ChIP-seq guidelines and practices of the ENCODE and modENCODE consortia

Landt, Stephen G.; Marinov, Georgi K.; Kundaje, Anshul; Kheradpour, Pouya; Pauli, Florencia; Batzoglou, Serafim; Bernstein, B; Bickel, Peter J.; Brown, James B.; Cayting, Philip; Chen, Yiwen; DeSalvo, Gilberto; Epstein, Charles B.; Fisher-Aylor, Katherine I.; Euskirchen, Ghia; Gerstein, Mark; Gertz, Jason; Hartemink, Alexander J.; Hoffman, Michael M.; Iyer, Vishwanath R.; Jung, Youngsook L.; Karmakar, Subhradip; Kellis, M; Kharchenko, Peter V.; Li, Qunhua; Liu, Tao; Liu, X. Shirley; Ma, Lijia; Milosavljevic, Aleksandar; Myers, R; Park, Peter J.; Pazin, Michael J.; Perry, Marc D.; Raha, Debasish; Reddy, Timothy E.; Rozowsky, Joel; Shoresh, Noam; Sidow, Arend; Slattery, Matthew; Stamatoyannopoulos, J; Tolstorukov, Michael Y.; White, Kevin P.; Xi, Simon; Farnham, Peggy J.; Lieb, Jason D.; Wold, B; Snyder, M

doi:10.1101/gr.136184.111

Cited by 1,829 publications

(1,980 citation statements)

References 51 publications

Supporting

Mentioning

1,873

Contrasting

Unclassified

Order By: Relevance

“…In vitro models accurately identify in vivo bound sequences We next evaluated DeepBind's performance using 506 in vivo ENCODE ChIP-seq data sets, which were preprocessed to remove protocol and laboratory biases 26 (Supplementary Table 4). Unlike experiments with in vitro data, these experiments were influenced by cell type-specific effects, transcription factor-nucleosome interactions, cooperation and competition between transcription factors and other cofactors, and pioneer transcription factors that can remodel chromatin and facilitate the binding of other transcription factors 27 .…”

Section: A N a Ly S I Smentioning

confidence: 99%

Predicting the sequence specificities of DNA- and RNA-binding proteins by deep learning

et al. 2015

View full text Add to dashboard Cite

Knowing the sequence specificities of DNA-and RNA-binding proteins is essential for developing models of the regulatory processes in biological systems and for identifying causal disease variants. Here we show that sequence specificities can be ascertained from experimental data with 'deep learning' techniques, which offer a scalable, flexible and unified computational approach for pattern discovery. Using a diverse array of experimental data and evaluation metrics, we find that deep learning outperforms other state-of-the-art methods, even when training on in vitro data and testing on in vivo data. We call this approach DeepBind and have built a stand-alone software tool that is fully automatic and handles millions of sequences per experiment. Specificities determined by DeepBind are readily visualized as a weighted ensemble of position weight matrices or as a 'mutation map' that indicates how variations affect binding within a specific sequence.DNA-and RNA-binding proteins play a central role in gene regulation, including transcription and alternative splicing. The sequence specificities of a protein are most commonly characterized using position weight matrices 1 (PWMs), which are easy to interpret and can be scanned over a genomic sequence to detect potential binding sites. However, growing evidence indicates that sequence specificities can be more accurately captured by more complex techniques 2-5 . Recently, 'deep learning' has achieved record-breaking performance in a variety of information technology applications 6,7 . We adapted deep learning methods to the task of predicting sequence specificities and found that they compete favorably with the state of the art. Our approach, called DeepBind, is based on deep convolutional neural networks and can discover new patterns even when the locations of patterns within sequences are unknown-a task for which traditional neural networks require an exorbitant amount of training data.There are several challenging aspects in learning models of sequence specificity using modern high-throughput technologies. First, the data come in qualitatively different forms. Protein binding microarrays (PBMs) 8 and RNAcompete assays 9 provide a specificity coefficient for each probe sequence, whereas chromatin immunoprecipitation (ChIP)-seq 10 provides a ranked list of putatively bound sequences of varying length, and HT-SELEX 11 generates a set of very high affinity sequences. Second, the quantity of data is large. A typical high-throughput experiment measures between 10,000 and 100,000 sequences, and it is computationally demanding to incorporate them all. Third, each data acquisition technology has its own artifacts, biases and limitations, and we must discover the pertinent specificities despite these unwanted effects. For example, ChIP-seq reads often localize to "hyper-ChIPable" regions of the genome near highly expressed genes 12 .DeepBind (Fig. 1) addresses the above challenges. (i) It can be applied to both microarray and sequencing data; (ii) it can learn from millions of...

show abstract

Section: A N a Ly S I Smentioning

confidence: 99%

Predicting the sequence specificities of DNA- and RNA-binding proteins by deep learning

et al. 2015

View full text Add to dashboard Cite

show abstract

“…However, the method is limited in its throughput by the need to create gene-specific antibodies or tagged transgenic lines, which can be technically challenging and expensive 8 . Alternative methods are needed for capturing genome-wide binding data for many organisms.…”

Section: Introductionmentioning

confidence: 99%

Mapping genome-wide transcription-factor binding sites using DAP-seq

Bartlett¹,

O’Malley²,

Huang³

et al. 2017

Nat Protoc

441

285

View full text Add to dashboard Cite

In order to enable low-cost, high-throughput generation of cistrome and epicistrome maps for any organism, we developed DNA affinity purification sequencing (DAP-seq), a transcription factor (TF) binding site discovery assay that couples affinity-purified TFs with next-generation sequencing of a genomic DNA library. The method is fast, inexpensive, and more easily scaled than ChIP-seq. DNA libraries are constructed using native genomic DNA from any source of interest, preserving cell-and tissue-specific chemical modifications that are known to impact TF binding (such as DNA methylation) and providing increased specificity compared to in silico motif prediction methods such as protein binding microarrays (PBM) and systematic evolution of ligands by exponential enrichment (SELEX). The resulting DNA library is incubated with an affinity-tagged in vitro expressed TF, and TF-DNA complexes are purified using magnetic separation of the affinity tag. Bound genomic DNA is eluted from the TF and sequenced using next-generation sequencing. Sequence reads are mapped to a reference genome, identifying genome-wide binding locations for each TF assayed, from which sequence motifs can then be

show abstract

“…S3A) is rather low. For example, the peaks reported for the HUVEC cell line on the ENCODE (Landt et al 2012) portal report in 37,920, 44,033, and 37,412 peaks called for the three laboratories, respectively, with 24.3% of regions reported by only one laboratory. Note that ENCODE uses the IDR algorithm (Li et al 2011) to select peaks that consistently appear on two replicates for each laboratory; thus, the number of peaks reported for each laboratory may be significantly affected by the quality of the replicates rather than by suboptimal performance from peak callers.…”

Section: Gc Affects Coverage and It Does So Differently In Differentmentioning

confidence: 99%

Accounting for GC-content bias reduces systematic errors and batch effects in ChIP-seq data

Teng

Irizarry

2017

Genome Res.

View full text Add to dashboard Cite

The main application of ChIP-seq technology is the detection of genomic regions that bind to a protein of interest. A large part of functional genomics' public catalogs is based on ChIP-seq data. These catalogs rely on peak calling algorithms that infer protein-binding sites by detecting genomic regions associated with more mapped reads (coverage) than expected by chance, as a result of the experimental protocol's lack of perfect specificity. We find that GC-content bias accounts for substantial variability in the observed coverage for ChIP-seq experiments and that this variability leads to false-positive peak calls. More concerning is that the GC effect varies across experiments, with the effect strong enough to result in a substantial number of peaks called differently when different laboratories perform experiments on the same cell line. However, accounting for GC content bias in ChIP-seq is challenging because the binding sites of interest tend to be more common in high GC-content regions, which confounds real biological signals with unwanted variability. To account for this challenge, we introduce a statistical approach that accounts for GC effects on both nonspecific noise and signal induced by the binding site. The method can be used to account for this bias in binding quantification as well to improve existing peak calling algorithms. We use this approach to show a reduction in false-positive peaks as well as improved consistency across laboratories.

show abstract

ChIP-seq guidelines and practices of the ENCODE and modENCODE consortia

Cited by 1,829 publications

References 51 publications

Predicting the sequence specificities of DNA- and RNA-binding proteins by deep learning

Predicting the sequence specificities of DNA- and RNA-binding proteins by deep learning

Mapping genome-wide transcription-factor binding sites using DAP-seq

Accounting for GC-content bias reduces systematic errors and batch effects in ChIP-seq data

Contact Info

Product

Resources

About