standOff. When we have changed our mind. Some entries in the table begin with a word in capitals and the reason for it. EVIDENCE CORRECTED — the finding still stands, but we had quoted the wrong sentence from the paper, and the quotation has been replaced with the right one. RETRACTED — the entry itself was wrong; it stays on the page, marked, with the reason, rather than being deleted. WITHDRAWN AND REPLACED — the entry was wrong, and a corrected version of it appears in the entries below. Each mark carries its own reason, in brackets, because the reasons differ from one entry to the next. An entry we have WITHDRAWN is removed from this page and kept in the record, with its reason. A line reading Audit and a date is a later re-reading of the entry against its whole sentence: DOUBTFUL means it was kept, and the reason says what is loose about it. Quotation marks inside the tagger's sentence. Where a tag is the tagger's own sentence, a phrase in quotation marks inside it is borrowed from the paper — the tagger's wording around it is not. The paper's own sentence is the evidence quote shown on the same row, so the borrowing can be checked against the thing borrowed from without leaving this page. Nothing inside quotation marks anywhere on this page is ours: if it is quoted, it is the paper's. Whose words a concept carries. A concept entry records a term the paper itself uses, so the term, the sense given for it, and its evidence are the paper's own words, quoted where the paper fixes them. This is the opposite of a sense entry in the tag table, which is the tagger's sentence about how the paper uses a term. Where a concept entry could not be grounded in the paper's words, the entry says so on its own row.| relation to other work | claim_relations | how this paper stands to another: agrees, contradicts, extends, qualifies |
| claim type | claim_type | the shape of the claim: existence, mechanism, method, critique, prediction, null |
| concept edge | concept_edges | an asserted relation between two concepts, with the sentence asserting it |
| concept | concepts | an idea the paper works with, quoted where the paper uses it |
| content tag | content_tags | a finding, definition, quantity, statistic or variable named in the paper |
| contribution | contribution | what the paper claims to add |
| control | controls | an experiment run to eliminate a rival explanation |
| limitation | limitations | what the paper says its evidence cannot support |
| mechanism | mechanisms | how the reported effect is said to come about |
| method | method | how the paper produced its evidence |
| object of study | object_of_study | what the paper is about — the thing it investigates |
| phenomenon | phenomenon | the effect or regularity under study |
| research question | research_question | the question the paper sets itself |
| sense | senses | a distinct use of a term, quoted where the paper fixes it |
| unrecordable relation | unhoused_relations | a relation the paper asserts that this vocabulary cannot express, with the reason |
The reference list is not part of the text we hold for this paper. Our copy was extracted from the publisher's page and the bibliography did not survive that extraction, so the article below ends before its references. Nothing has been removed by us.
Abstract. Functional genomics assays based on high-throughput sequencing greatly expand our ability to understand the genome. object_of_studyclaim_typeconcepts↗sensescontent_tagscontent_tags↗Here, we define the ENCODE blacklist- a comprehensive set of regions in the human, mouse, worm, and fly genomes that have anomalous, unstructured, or high signal in next-generation sequencing experiments independent of cell line or experiment. The removal of the ENCODE blacklist is an essential quality measure when analyzing functional genomics data. Introduction concepts↗concepts↗concepts↗concepts↗concepts↗concepts↗concepts↗concepts↗concepts↗concepts↗concepts↗concepts↗concepts↗concepts↗unhoused_relationssensesconcept_edgesconcept_edgesThe coupling of high-throughput technology with classic genomics assays enables us to study genome-wide architecture and regulation. Assays using high-throughput sequencing as a read-out of a genomic signal rely on an accurate genomic annotation and mapping. Inconsistencies in the underlying annotation exist at regions where assembly has been difficult. For instance, repetitive regions may be collapsed or under-represented in the reference sequence relative to the actual underlying genomic sequence. Resulting analysis of these regions can lead to inaccurate interpretation, as there may be significant enrichment of signal because of amplification of noise 1 , 2 . sensesclaim_relationscontent_tagsProblematic regions such as these have generally been ignored and unfiltered because they have not been found to affect the signal in the final analyses. However, in functional genomics assays such as chromatin immunoprecipitation followed by genome sequencing (ChIP-seq), accuracy in peak calling and downstream analyses is essential. Alignments in these problematic regions should be identified and filtered before application of any threshold, normalization, or peak calling as they can dramatically bias the results 2 . concepts↗concepts↗concepts↗concepts↗concepts↗concepts↗concepts↗concepts↗concepts↗The use of exclusive regions of “blacklists”, or regions where genome assembly results in erroneous signal, to remove signal-artifact regions in ChIP-seq experiments has been employed throughout the ENCODE project production phase 1 , 3 , 4 . concepts↗concepts↗sensesclaim_relationsconcept_edgescontent_tagsThe original ENCODE blacklist, termed the Duke Excluded Regions (DER), was manually curated on the Homo sapiens (human) genome assembly GRCh37 (hereafter referred to as hg19) to cover a large number of repeat elements in the genome, particularly rRNA, alpha satellites, and other simple repeats. This list was further updated, now referred to as ENCODE Data Analysis Center (DAC) blacklisted regions, to include regions of high signal that presumably represent unannotated repeats in the genome. The removal of these regions eliminated significant background noise that otherwise would have been thought to have been due to biological variation 2 . While this list was comprehensive, a significant amount of manual annotation was required to generate the final set of regions that would be laborious to apply to updated builds. The affected regions were broad, covering on average 45 kb with the largest being 1.4 Mb. Additionally, artifact regions are not human genome specific, and there was a need for identification of organism-specific regions. concepts↗concepts↗concepts↗concepts↗concepts↗concepts↗concepts↗concepts↗concepts↗concepts↗concepts↗concepts↗concepts↗concepts↗concepts↗concepts↗unhoused_relationssensessensessensesTo generate blacklists in an objective and systematic manner, we developed an automatic procedure to flag regions that appear to have artifact signal. Regions are flagged using uniform criteria applied across a large number of samples. All ENCODE, mouse ENCODE, and modENCODE input ChIP-seq samples (control data for ChIP-seq) were used for ENCODE ( Homo sapiens : hg19 and the updated assembly GRCh38/hg38), mouse ENCODE ( Mus musculus: mm9 and mm10), and modENCODE ( Caenorhabditis elegans: ce10 and ce11, Drosophila melanogaster: dm3 and dm6) analyses, respectively. To identify regions for inspection, our method searches for regions that provide the signature of existing in multiple copies and are thus overrepresented in control “input” sequences. These “input” datasets were generated as controls for ChIP-seq experiments using randomly sheared DNA regions from non-immunoprecipitated chromatin. We examined all 1 kb windows with 100 bp overlap to identify such regions. Input samples are scored with input read depth and mappability, quantile normalized, and the median signal is selected (See methods). This defines a comprehensive and cell-type agnostic signal across the genome that is unaffected by high signal from a particular cell-line (eg. CNVs) or low signal due to differential processing of input data. research_questioncontributionmethod↗limitationscontrolscontrolsconcepts↗concepts↗concepts↗concepts↗sensesconcept_edgescontent_tagscontent_tags↗content_tagsRegions with read depths or multi-mapping read rates in the top 1% are considered likely artifacts (Supplemental Fig. 1 ). In all cases, the mitochondrial DNA and any reads mapping to these sequences are pre-filtered from analysis and are considered part of the blacklist. A blacklist was built for the human, mouse, worm, and fly genomes using all reads from input samples. concepts↗concepts↗concepts↗concepts↗In each organism, only a small portion of the genome was flagged as containing artifact sequence signal (Figs 1 and 2a ). However, these regions were enriched for ChIP-seq reads in ChIP experiment for transcription factors (Fig. 1 ) and are particularly enriched for reads and peaks from lower quality experiments. concepts↗content_tagscontent_tagscontent_tags↗In fact, ENCODE uses this as a quality control metric with some experiments having up to 87% of reads falling into blacklisted regions 5 . In Fig. 1 we show the distribution of all input reads mapped across chromosome 1, where reads mapping to blacklist regions are represented in red. phenomenoncontent_tagsThe signal at blacklist regions are extremely high even though they account for a small fraction of the mappable chromosome (Fig. 1b ). For example, this represented 582 million of 2.5 billion uniquely aligning reads mapping to blacklisted regions in the human ENCODE ChIP-seq data in hg19. These findings emphasize the extreme nature of these artifact regions and highlight the importance of filtering these regions to avoid incorrect biological conclusions. Figure 1 Blacklist regions are tightly distributed across the chromosome and sequester high read mapping signals. ( a ) Distribution of mapped reads along human chromosome 1 in hg19. ( b ) An example blacklisted region on chromosome 1. Displayed are pre-filtered ENCODE ChIP-seq peak calls, quantile normalized median read signal (Reads), and quantile normalized median multimapped read signal (Multi). Axes are scaled for illustrative purposes and signal values are truncated at approximately 10-fold enrichment. Signal in these regions are up to 6400× background levels. ( c ) An example “normal” ENCODE ChIP-seq peak region on chromosome 1 selected as a region containing ChIP-seq peaks. Figure 2 Blacklist regions account for a significant portion of ChIP-seq reads, are driven by artifacts in genome assemblies, and removal of these regions is essential to removing noise in genomics assays. ( a ) The number of blacklisted regions across species with their average size, genomic coverage, and input datasets excluding assembly gaps used for hg38, mm10, dm6, and ce11 respectively. ( b ) An UpSet plot displaying the breakdown of uniquely annotated regions in hg19 and hg38, and the shared regions between them. Low-mappability (Low-Map.) regions account for the majority of unique regions in both hg19 and hg38. ( c ) Applying the blacklist to ChIP-seq peaks results in an overall reduced correlation and, in the highlighted example, results in a more biologically meaningful interpretation of the data. concepts↗concepts↗concepts↗concepts↗concepts↗concepts↗sensesconcept_edgesWe investigated the underlying characteristics of our automated hg19 blacklisted regions and their agreement with previously published lists 6 , which included our manually curated hg19 blacklist (DAC) (Supplemental Fig. 2 ). Though satellite repeats were used in the original ENCODE blacklists, they represent a small portion of the automated hg19 blacklist and are generally repeated in the genome annotation (Supplemental Fig. 2c ). Additionally, they are not uniquely mapped with the algorithms used which prevents alignments to these regions. The automated hg19 and DAC blacklists detect the vast majority of regions flagged by a similar and complementary technique 6 (Supplemental Fig. 2b ). Of the regions unique to the automated hg19 blacklist, all cover gaps in the genome assembly (Supplemental Fig. 2c ), lending evidence that these regions of the genome are incomplete in the hg19 assembly. Indeed, a large number of these regions were patched in the next iteration of hg19 or dropped in the GRCh38 assembly (Supplemental Fig. 2c ). limitationscontent_tagsAlmost all of the flagged regions also included nuclear mitochondrial DNA segments (NUMTs, Supplemental Fig. 2c ), a criterion that was overlooked in the initial manual blacklists. There are many mitochondrial genomes in comparison to the singular nuclear genome, leading to a high read depth of NUMTs. Additionally, NUMT sequences are scattered throughout the genome, contributing to overrepresentation of NUMTs in the input sequence. For these reasons, it is critical to include these sequences in the blacklist. A majority of the regions that were flagged by the DAC blacklist but missed in the automated hg19 blacklist were defined by repeatmasker class annotations as Satellite repeats (Supplemental Fig. 2d ). While many of these repeat regions contain anomalous signal, those that were excluded from the automated hg19 blacklist do not show high signal and are uniquely mapped in hg19. None of the regions unique to the DAC blacklist were patched or removed in the new assembly. concepts↗concepts↗concepts↗concepts↗sensesWe next sought to characterize the differences between regions blacklisted from the automated pipeline in hg19 compared to the blacklist from the hg38 genome assembly. Generally, the same classes of regions are enriched in both assemblies. Many regions do not overlap due to assembly differences such as the expanded centromere and satellite sequences that are features of the hg38 assembly as well as fixed/new gaps that vary in both builds (Fig. 2b ). A large portion of the differences occurs at low-mappability (Low Map.) annotations which represent short repetitive elements in the genome assembly that are poorly mappable and as a consequence do not map well between assemblies (Fig. 2b ). mechanismscontent_tagsOverall, these differences lead to the conclusion that the major differences between blacklists are due to underlying changes in the sequence assemblies. Consequently, this lends to the hypothesis that the driving factor behind artifact regions in the genome are due to issues in the assembly rather than other factors. concepts↗concepts↗concepts↗concepts↗concepts↗concepts↗concepts↗sensesconcept_edgesFinally, to demonstrate the artificial correlation created by these peak regions, we performed a correlation analysis of the ENCODE peak regions in the human genome using a blacklist-filtered set as well as an unfiltered set of peaks (Fig. 2c ). mechanismscontent_tagsIn the unfiltered set of peaks, blacklist regions sequester a large portion of ChIP-seq reads, leading to an illusion of high correlation of these regions with others. After blacklist filtering, the correlation structure is more distinct. controlsconcepts↗content_tagsAs a specific example, the correlation of REST (a known repressor) auto correlates with other TFs (most of them activating) without the filtering of the blacklist. The removal of the blacklist regions removes spurious correlations seen with REST has essentially disappeared as would be expected given the known biology. We highlight the clear case of removing the noise in REST correlation, but the same standard holds true for the remaining factors in Fig. 2c . concepts↗concepts↗concepts↗concepts↗concept_edgesThe ENCODE blacklists have been used to filter all of the ChIP-seq data from the ENCODE project and improvements in data from the application of the blacklist to these data are a key evaluation metric used by the consortium. For a complete list of artifact effects on peaks from all ChIP experiments used in ENCODE, we have provided a reference to the ENCODE quality control metric spreadsheet 5 . Furthermore, another detailed analysis of the detrimental effects of not excluding these artifact regions has been previously described 2 . Biological validations of the most robust signal regions will likely result in testing of these artifact regions, potentially resulting in incorrect biological conclusions. Therefore, identification of these regions and subsequent filtering lead to more accurate and stable results across experiments. concepts↗concepts↗concepts↗concepts↗concepts↗concepts↗concepts↗unhoused_relationssensesThe method implemented here requires a significant amount of input sequencing data from different sources in order to generate an accurate blacklist. For our analyses, we use all available ENCODE ChIP-seq input datasets to estimate the genomic regions that have these artifact properties, and the use of multiple cell-types is important to avoid blacklisting regions that are specific to a single cell-type or tissue. limitationscontent_tagsWe also caution that the blacklists are specific to each genome assembly and a lift over from an old assembly is not meaningful or valid. limitationscontent_tagscontent_tagsFinally, those studying genes in unmappable regions of the genome will find their data filtered by the blacklist. These regions account for ~3% of human protein coding genes that have previously been shown to be unmappable using short-read technologies 7 . concepts↗concepts↗concepts↗concepts↗concepts↗concepts↗concepts↗concepts↗concepts↗concepts↗concepts↗concepts↗concepts↗concepts↗unhoused_relationsunhoused_relationssensessensesconcept_edgesWe present a resource of genomic regions that should be identified and either filtered from study or analyzed independently for better understanding as to their potential regulatory function. limitationsIt is important to note that we do not propose a single blacklist that can encapsulate error defined across all NGS based assays. The presented blacklist shows high concordance between chromatin-based filtering (DNase/ATAC-DAC blacklist) and ChIP-seq input based filtering. This is not surprising given that the input DNA for ChIP-seq has been shown to be a proxy for lightly digested open chromatin assays 8 . However, the same criteria cannot be applied for whole genome sequencing (WGS) filtering and RNA-seq filtering. WGS filtering does not result in poorer annotations if there are higher read depths in regions, and therefore this method would be counterproductive to genome assembly. In the case of RNA-seq, more cell-type specific corrections for copy number are appropriate as there is virtually no overlap of coding regions with existing blacklist regions. The method presented is employed by the ENCODE project, as well as many other established analysis pipelines, and allows for a noise filtering on DNase-seq, ATAC-seq, and ChIP-seq datasets to help improve the accuracy of studies using these data. The removal of blacklists differs from the typical removal of signals from duplicate reads since these regions are problematic across different cell types and individual experiments. Ultimately, the removal of blacklists should be integrated within genomic assay analysis pipelines that incorporate high-throughput sequencing in order to assess biologically relevant and true signals. Methods Selection of input datasets concepts↗concepts↗concepts↗All data were acquired from the ENCODE Data Coordination Center. Using a previously published perl script ( https://github.com/Boyle-Lab/ENCODE-API-Apps ) 9 , we queried the ENCODE DCC API for unfiltered bam files labeled “input” that were released for the correct genome assembly. In the case of humans, these bam files were merged based on ENCODE assigned donor accession numbers to collapse data by cell type or individual. This was performed using ‘samtools sort’ to first sort all samples, ‘samtools merge’ to merge, and finally ‘samtools index’ to generate a new index of the resulting bam files 10 . Generation of mappability data concepts↗concepts↗unhoused_relationsThe Umap tool was used to identify all positions on both strands of a target genome for which reads of a desired length starting at that position are uniquely mappable 11 . Building the blacklist concepts↗concepts↗concepts↗concepts↗concepts↗concepts↗concepts↗sensesFor each input dataset (or merged input for human) from ENCODE, the number of reads per mappable base and the number of multimapping reads per million reads is calculated for each bin of 1 kb with 100 bp overlap across all chromosomes. The values across bins are then quantile normalized and a standard value at the 50% quantile is selected to represent each bin. This threshold was selected to avoid high signal outliers from individual cell types (for example, from copy number variants) and to avoid low signal from failed or incorrectly labeled input datasets. The standard values across the genome are then flagged if they are in the top 0.1% of signal for either read depth or mappability. Neighboring regions are merged if they maintain a signal in the top 1% of all signal or if they have no signal due to no mappability in the genome and any flagged regions within 20 kb were combined. This generates contiguous regions of abnormal signal across the genome. Supplementary information SUPPLEMENTARY INFO
Relations on this page carry a quotation for the claim. Where a relation was separately evidenced — and so tested against the vocabulary for its type — the row says so; 7 of 8 relations here carry that mark. The rest are the tagger’s reading of the sentence quoted beside them.
| facet | tag | field | conf. | whose words |
|---|---|---|---|---|
| object_of_study | A PUBLISHED LIST OF WHAT DOES NOT COUNT - genomic regions excluded from every ENCODE analysis by convention, and now from many pipelines outside it | object_of_study | the paper's words | |
| research_question | which regions of a genome produce signal that is an artefact of the assembly rather than of the sample, and how to identify them without hand curation | research_question | the paper's words | |
| contribution | an automated, threshold-based blacklist for human, mouse, worm and fly, replacing a hand-curated list - and a demonstration that applying it changes published correlations | contribution | the paper's words | |
| claim_type | method | claim_type | the paper's words | |
| phenomenon | regions where the reference assembly is wrong or incomplete, which collect reads at up to 6,400x background and are then read as biology | phenomenon | the paper's words | |
| method | system-or-tool-building also in: system-or-tool-building | method | the paper's words | |
| mechanisms | THE ARTEFACT IS A PROPERTY OF THE REFERENCE, NOT OF THE SAMPLE: `the major differences between blacklists are due to underlying changes in the sequence assemblies ... the driving factor behind artifact regions in the genome are due to issues in the assembly rather than other factors`. Change the reference and the exclusion list changes | mechanisms[0] | 0.95 | the tagger's words |
| mechanisms | and the failure mode is an ILLUSION OF STRUCTURE: `blacklist regions sequester a large portion of ChIP-seq reads, leading to an illusion of high correlation of these regions with others. After blacklist filtering, the correlation structure is more distinct`. The artefact does not add noise, it adds AGREEMENT | mechanisms[1] | 0.95 | the tagger's words |
| limitations | THE LIST REMOVES ~3% OF HUMAN PROTEIN CODING GENES FROM ANALYSIS, AND THE PAPER SAYS SO: `those studying genes in unmappable regions of the genome will find their data filtered by the blacklist`. A researcher whose gene is on the list cannot study it with these pipelines | limitations[0] | 0.95 | the tagger's words |
| limitations | IT IS BOUND TO ONE REFERENCE VERSION: `the blacklists are specific to each genome assembly and a lift over from an old assembly is not meaningful or valid`. The list is not a fact about the genome; it is a fact about a build of the genome | limitations[1] | 0.95 | the tagger's words |
| limitations | and it is ASSAY-SPECIFIC, declared against the paper's own generality: `we do not propose a single blacklist that can encapsulate error defined across all NGS based assays ... the same criteria cannot be applied for whole genome sequencing filtering and RNA-seq filtering` | limitations[2] | 0.95 | the tagger's words |
| limitations | the threshold is A CHOICE STATED AS ONE: regions in `the top 1%` of read depth or multi-mapping rate `are considered LIKELY artifacts`. Nothing in the paper derives 1%; it is where the line was put | limitations[3] | 0.95 | the tagger's words |
| limitations | and the list it replaces MISSED AN ENTIRE CRITERION: nuclear mitochondrial DNA segments were `a criterion that was overlooked in the initial manual blacklists`. The hand-curated version had been in production use for years | limitations[4] | 0.95 | the tagger's words |
| controls | the blacklist was built from INPUT samples - randomly sheared, non-immunoprecipitated chromatin - across many cell types, quantile normalised, median taken eliminates: cell-line-specific signal (copy number variants) and processing differences as sources of the flagged regions; the paper says the multiple cell types are there for exactly that · outcome: held | controls[0] | 0.95 | the tagger's words |
| controls | whether applying the list changes conclusions, tested on a named case eliminates: the objection that filtering is cosmetic - REST, a known repressor, auto-correlates with activating factors UNTIL the blacklist is applied, and the spurious correlation 'has essentially disappeared as would be expected given the known biology' · outcome: held | controls[1] | 0.95 | the tagger's words |
| controls | whether the 1% threshold is derived eliminates: nothing - regions in 'the top 1%' are 'considered LIKELY artifacts', and no derivation, sensitivity analysis or alternative cut is offered · outcome: failed | controls[2] | 0.95 | the tagger's words |
| concepts | ENCODE blacklist also in: ENCODE blacklist | concepts[0] | 0.95 | the paper's words |
| concepts | Duke Excluded Regions also in: Duke Excluded Regions | concepts[1] | 0.95 | the paper's words |
| concepts | ENCODE Data Analysis Center (DAC) blacklisted regions | concepts[2] | 0.95 | the paper's words |
| concepts | artifacts also in: artifacts | concepts[3] | 0.95 | the paper's words |
| concepts | quality control metric also in: quality control metric | concepts[4] | 0.95 | the paper's words |
| concepts | REST also in: REST | concepts[5] | 0.95 | the paper's words |
| concepts | high-throughput technology also in: high-throughput technology | concepts[6] | 0.95 | the paper's words |
| concepts | classic genomics assays also in: classic genomics assays | concepts[7] | 0.95 | the paper's words |
| concepts | genome-wide architecture and regulation | concepts[8] | 0.95 | the paper's words |
| concepts | accurate genomic annotation and mapping | concepts[9] | 0.95 | the paper's words |
| concepts | inconsistencies in the underlying annotation | concepts[10] | 0.95 | the paper's words |
| concepts | repetitive regions also in: repetitive regions | concepts[11] | 0.95 | the paper's words |
| concepts | collapsed or under-represented also in: collapsed or under-represented | concepts[12] | 0.95 | the paper's words |
| concepts | reference sequence also in: reference sequence | concepts[13] | 0.95 | the paper's words |
| concepts | inaccurate interpretation also in: inaccurate interpretation | concepts[14] | 0.95 | the paper's words |
| concepts | significant enrichment of signal also in: significant enrichment of signal | concepts[15] | 0.95 | the paper's words |
| concepts | amplification of noise also in: amplification of noise | concepts[16] | 0.95 | the paper's words |
| concepts | chromatin immunoprecipitation followed by genome sequencing (ChIP-seq) | concepts[17] | 0.95 | the paper's words |
| concepts | peak calling also in: peak calling | concepts[18] | 0.95 | the paper's words |
| concepts | downstream analyses also in: downstream analyses | concepts[19] | 0.95 | the paper's words |
| concepts | exclusive regions of “blacklists” also in: exclusive regions of “blacklists” | concepts[20] | 0.95 | the paper's words |
| concepts | signal-artifact regions also in: signal-artifact regions | concepts[21] | 0.95 | the paper's words |
| concepts | rRNA also in: rRNA | concepts[22] | 0.95 | the paper's words |
| concepts | alpha satellites also in: alpha satellites | concepts[23] | 0.95 | the paper's words |
| concepts | unannotated repeats also in: unannotated repeats | concepts[24] | 0.95 | the paper's words |
| concepts | significant background noise also in: significant background noise | concepts[25] | 0.95 | the paper's words |
| concepts | biological variation also in: biological variation | concepts[26] | 0.95 | the paper's words |
| concepts | manual annotation also in: manual annotation | concepts[27] | 0.95 | the paper's words |
| concepts | organism-specific regions also in: organism-specific regions | concepts[28] | 0.95 | the paper's words |
| concepts | automatic procedure also in: automatic procedure | concepts[29] | 0.95 | the paper's words |
| concepts | uniform criteria also in: uniform criteria | concepts[30] | 0.95 | the paper's words |
| concepts | input ChIP-seq samples also in: input ChIP-seq samples | concepts[31] | 0.95 | the paper's words |
| concepts | control data for ChIP-seq also in: control data for ChIP-seq | concepts[32] | 0.95 | the paper's words |
| concepts | signature of existing in multiple copies | concepts[33] | 0.95 | the paper's words |
| concepts | control “input” sequences also in: control “input” sequences | concepts[34] | 0.95 | the paper's words |
| concepts | randomly sheared DNA regions also in: randomly sheared DNA regions | concepts[35] | 0.95 | the paper's words |
| concepts | non-immunoprecipitated chromatin also in: non-immunoprecipitated chromatin | concepts[36] | 0.95 | the paper's words |
| concepts | 1 kb windows also in: 1 kb windows | concepts[37] | 0.95 | the paper's words |
| concepts | 100 bp overlap also in: 100 bp overlap | concepts[38] | 0.95 | the paper's words |
| concepts | input read depth also in: input read depth | concepts[39] | 0.95 | the paper's words |
| concepts | mappability also in: mappability | concepts[40] | 0.95 | the paper's words |
| concepts | quantile normalized also in: quantile normalized | concepts[41] | 0.95 | the paper's words |
| concepts | median signal also in: median signal | concepts[42] | 0.95 | the paper's words |
| concepts | cell-type agnostic signal also in: cell-type agnostic signal | concepts[43] | 0.95 | the paper's words |
| concepts | copy number variants also in: copy number variants | concepts[44] | 0.95 | the paper's words |
| concepts | multi-mapping read rates also in: multi-mapping read rates | concepts[45] | 0.95 | the paper's words |
| concepts | mitochondrial DNA also in: mitochondrial DNA | concepts[46] | 0.95 | the paper's words |
| concepts | reads mapping to these sequences also in: reads mapping to these sequences | concepts[47] | 0.95 | the paper's words |
| concepts | small portion of the genome also in: small portion of the genome | concepts[48] | 0.95 | the paper's words |
| concepts | lower quality experiments also in: lower quality experiments | concepts[49] | 0.95 | the paper's words |
| concepts | small fraction of the mappable chromosome | concepts[50] | 0.95 | the paper's words |
| concepts | incorrect biological conclusions also in: incorrect biological conclusions | concepts[51] | 0.95 | the paper's words |
| concepts | automated hg19 blacklist also in: automated hg19 blacklist | concepts[52] | 0.95 | the paper's words |
| concepts | gaps in the genome assembly also in: gaps in the genome assembly | concepts[53] | 0.95 | the paper's words |
| concepts | nuclear mitochondrial DNA segments also in: nuclear mitochondrial DNA segments | concepts[54] | 0.95 | the paper's words |
| concepts | high read depth of NUMTs also in: high read depth of NUMTs | concepts[55] | 0.95 | the paper's words |
| concepts | many mitochondrial genomes also in: many mitochondrial genomes | concepts[56] | 0.95 | the paper's words |
| concepts | overrepresentation of NUMTs also in: overrepresentation of NUMTs | concepts[57] | 0.95 | the paper's words |
| concepts | assembly differences also in: assembly differences | concepts[58] | 0.95 | the paper's words |
| concepts | low-mappability (Low Map.) annotations | concepts[59] | 0.95 | the paper's words |
| concepts | major differences between blacklists also in: major differences between blacklists | concepts[60] | 0.95 | the paper's words |
| concepts | issues in the assembly also in: issues in the assembly | concepts[61] | 0.95 | the paper's words |
| concepts | artificial correlation also in: artificial correlation | concepts[62] | 0.95 | the paper's words |
| concepts | blacklist regions also in: blacklist regions | concepts[63] | 0.95 | the paper's words |
| concepts | blacklist-filtered set also in: blacklist-filtered set | concepts[64] | 0.95 | the paper's words |
| concepts | unfiltered set of peaks also in: unfiltered set of peaks | concepts[65] | 0.95 | the paper's words |
| concepts | illusion of high correlation also in: illusion of high correlation | concepts[66] | 0.95 | the paper's words |
| concepts | more distinct also in: more distinct | concepts[67] | 0.95 | the paper's words |
| concepts | spurious correlations also in: spurious correlations | concepts[68] | 0.95 | the paper's words |
| concepts | key evaluation metric also in: key evaluation metric | concepts[69] | 0.95 | the paper's words |
| concepts | biological validations also in: biological validations | concepts[70] | 0.95 | the paper's words |
| concepts | more accurate and stable results across experiments | concepts[71] | 0.95 | the paper's words |
| concepts | identification of these regions and subsequent filtering | concepts[72] | 0.95 | the paper's words |
| concepts | method implemented here also in: method implemented here | concepts[73] | 0.95 | the paper's words |
| concepts | significant amount of input sequencing data from different sources | concepts[74] | 0.95 | the paper's words |
| concepts | multiple cell-types also in: multiple cell-types | concepts[75] | 0.95 | the paper's words |
| concepts | specific to each genome assembly also in: specific to each genome assembly | concepts[76] | 0.95 | the paper's words |
| concepts | lift over from an old assembly also in: lift over from an old assembly | concepts[77] | 0.95 | the paper's words |
| concepts | unmappable regions of the genome also in: unmappable regions of the genome | concepts[78] | 0.95 | the paper's words |
| concepts | short-read technologies also in: short-read technologies | concepts[79] | 0.95 | the paper's words |
| concepts | resource of genomic regions also in: resource of genomic regions | concepts[80] | 0.95 | the paper's words |
| concepts | single blacklist also in: single blacklist | concepts[81] | 0.95 | the paper's words |
| concepts | all NGS based assays also in: all NGS based assays | concepts[82] | 0.95 | the paper's words |
| concepts | chromatin-based filtering also in: chromatin-based filtering | concepts[83] | 0.95 | the paper's words |
| concepts | ChIP-seq input based filtering also in: ChIP-seq input based filtering | concepts[84] | 0.95 | the paper's words |
| concepts | lightly digested open chromatin assays | concepts[85] | 0.95 | the paper's words |
| concepts | input DNA for ChIP-seq also in: input DNA for ChIP-seq | concepts[86] | 0.95 | the paper's words |
| concepts | whole genome sequencing (WGS) filtering | concepts[87] | 0.95 | the paper's words |
| concepts | RNA-seq filtering also in: RNA-seq filtering | concepts[88] | 0.95 | the paper's words |
| concepts | DNase-seq, ATAC-seq, and ChIP-seq datasets | concepts[89] | 0.95 | the paper's words |
| concepts | typical removal of signals from duplicate reads | concepts[90] | 0.95 | the paper's words |
| concepts | removal of blacklists also in: removal of blacklists | concepts[91] | 0.95 | the paper's words |
| concepts | genomic assay analysis pipelines also in: genomic assay analysis pipelines | concepts[92] | 0.95 | the paper's words |
| concepts | biologically relevant and true signals | concepts[93] | 0.95 | the paper's words |
| concepts | ENCODE Data Coordination Center also in: ENCODE Data Coordination Center | concepts[94] | 0.95 | the paper's words |
| concepts | ENCODE DCC API also in: ENCODE DCC API | concepts[95] | 0.95 | the paper's words |
| concepts | donor accession numbers also in: donor accession numbers | concepts[96] | 0.95 | the paper's words |
| concepts | Umap tool also in: Umap tool | concepts[97] | 0.95 | the paper's words |
| concepts | both strands of a target genome also in: both strands of a target genome | concepts[98] | 0.95 | the paper's words |
| concepts | number of reads per mappable base also in: number of reads per mappable base | concepts[99] | 0.95 | the paper's words |
| concepts | number of multimapping reads per million reads | concepts[100] | 0.95 | the paper's words |
| concepts | 50% quantile also in: 50% quantile | concepts[101] | 0.95 | the paper's words |
| concepts | top 0.1% of signal also in: top 0.1% of signal | concepts[102] | 0.95 | the paper's words |
| concepts | top 1% of all signal also in: top 1% of all signal | concepts[103] | 0.95 | the paper's words |
| concepts | within 20 kb also in: within 20 kb | concepts[104] | 0.95 | the paper's words |
| concepts | contiguous regions of abnormal signal | concepts[105] | 0.95 | the paper's words |
| unhoused_relations | high-throughput technology → genome-wide architecture and regulation Coupling high-throughput technology with classic assays enables study of genome-wide architecture and regulation. Enables-study-of is absent from the eight types. | unhoused_relations[0] | 0.95 | the tagger's words |
| unhoused_relations | automatic procedure → uniform criteria The automatic procedure flags regions using uniform criteria across many samples. Implements/uses-rule is absent from the governed edge types. | unhoused_relations[1] | 0.95 | the tagger's words |
| unhoused_relations | Umap tool → both strands of a target genome Umap is used to identify uniquely mappable read starts on both genome strands. Operates-on/identifies-in is absent from the eight types. | unhoused_relations[2] | 0.95 | the tagger's words |
| unhoused_relations | method implemented here → significant amount of input sequencing data from different sources The blacklist method requires substantial input data from diverse sources. Requires is absent from the eight types. | unhoused_relations[3] | 0.95 | the tagger's words |
| unhoused_relations | input DNA for ChIP-seq → lightly digested open chromatin assays ChIP-seq input DNA is reported as a proxy for lightly digested open-chromatin assays. Proxy-for is absent from the edge registry. | unhoused_relations[4] | 0.95 | the tagger's words |
| unhoused_relations | single blacklist → all NGS based assays The authors reject a single blacklist covering error across every NGS-based assay. Applicable-across is absent, and same-as or part-of would be false. | unhoused_relations[5] | 0.95 | the tagger's words |
| senses | `blacklist` - A PUBLISHED CONTROLLED LIST OF WHAT DOES NOT COUNT, and removing it is `an essential quality measure`. Not a recommendation: ENCODE filters every ChIP-seq dataset with it and uses the proportion of reads it catches as A QUALITY CONTROL METRIC FOR THE EXPERIMENT | senses[0] | 0.95 | the tagger's words |
| senses | `artifact region` - OPERATIONALLY, A 1 kb WINDOW IN THE TOP 1% of input read depth or multi-mapping rate, after quantile normalisation, across many cell types. The class is defined by a percentile, and the percentile is not derived from anything | senses[1] | 0.95 | the tagger's words |
| senses | `Duke Excluded Regions` -> `ENCODE Data Analysis Center blacklisted regions` - THE SAME OBJECT UNDER TWO NAMES, three sentences apart, with both naming verbs present: `TERMED the Duke Excluded Regions (DER)` and `now referred to as ENCODE Data Analysis Center (DAC) blacklisted regions` | senses[2] | 0.95 | the tagger's words |
| senses | what changed about the field: these regions `have generally been ignored and unfiltered because they have not been found to affect the signal in the final analyses` - and the paper's case is that in ChIP-seq they `can dramatically bias the results`. The same regions, a different assay, an opposite verdict | senses[3] | 0.95 | the tagger's words |
| senses | Assembly artifact - collapsed or underrepresented repeats can amplify noise into apparent genomic enrichment | senses[4] | 0.95 | the tagger's words |
| senses | Blacklist production shift - laborious human annotation replaced by an objective uniform procedure across many input samples | senses[5] | 0.95 | the tagger's words |
| senses | ChIP-seq input control - randomly sheared non-immunoprecipitated chromatin used to expose multi-copy regions | senses[6] | 0.95 | the tagger's words |
| senses | Cell-type-agnostic score - quantile-normalized median read depth and mappability across controls | senses[7] | 0.95 | the tagger's words |
| senses | NUMT artifact - many mitochondrial genomes plus dispersed nuclear inserts yield high input read depth | senses[8] | 0.95 | the tagger's words |
| senses | Build-specific blacklist - changed centromeres, satellites, gaps, and mappability make old-build lift-over invalid | senses[9] | 0.95 | the tagger's words |
| senses | Unfiltered correlation - blacklist regions sequester reads and create an illusion of factor similarity | senses[10] | 0.95 | the tagger's words |
| senses | Input diversity requirement - multiple sources and cell types prevent a single biological state from being blacklisted | senses[11] | 0.95 | the tagger's words |
| senses | Assay scope - chromatin-accessibility and ChIP-seq filtering transfer, but the same criteria are inappropriate for WGS and RNA-seq | senses[12] | 0.95 | the tagger's words |
| senses | Blacklist versus duplicate-read filtering - cross-cell-type region exclusion rather than experiment-local duplicate removal | senses[13] | 0.95 | the tagger's words |
| senses | Blacklist construction - 1 kb overlapping bins, normalized median statistics, extreme-signal flags, and distance-based merging | senses[14] | 0.95 | the tagger's words |
| claim_relations | extends → the hand-curated Duke Excluded Regions and DAC blacklists used throughout the ENCODE production phase - automated here, and extended to mouse, worm and fly | claim_relations[0] | 0.95 | the tagger's words |
| claim_relations | contradicts → the prior practice of leaving these regions unfiltered, which rested on their 'not having been found to affect the signal in the final analyses' | claim_relations[1] | 0.95 | the tagger's words |
| concept_edges | Duke Excluded Regions —renamed-across-fields→ ENCODE Data Analysis Center (DAC) blacklisted regions Audit 2026-09-28: DOUBTFUL - read against its span and kept, but the blacklist was renamed within the ENCODE project; renamed-across-fields overstates. | concept_edges[0] | 0.95 | the tagger's words |
| concept_edges | repetitive regions —causes→ inaccurate interpretation the relation itself is separately evidenced · strength of the relation: 0.7 · note: Analysis of collapsed or underrepresented repeated sequence can lead to inaccurate interpretation. also in: repetitive regions · inaccurate interpretation | concept_edges[1] | 0.95 | the tagger's words |
| concept_edges | amplification of noise —causes→ significant enrichment of signal the relation itself is separately evidenced · strength of the relation: 0.7 · note: The paper says apparent enrichment may arise because noise is amplified. | concept_edges[2] | 0.95 | the tagger's words |
| concept_edges | mitochondrial DNA —part-of→ ENCODE blacklist the relation itself is separately evidenced · strength of the relation: 1 · note: Mitochondrial DNA is pre-filtered and explicitly considered part of the blacklist.
Audit 2026-09-28: DOUBTFUL - read against its full span and kept, but the span writes 'the blacklist', not 'ENCODE blacklist'. also in: mitochondrial DNA · ENCODE blacklist | concept_edges[3] | 0.95 | the tagger's words |
| concept_edges | many mitochondrial genomes —causes→ high read depth of NUMTs the relation itself is separately evidenced · strength of the relation: 1 · note: The many mitochondrial copies are stated to lead to high NUMT read depth. | concept_edges[4] | 0.95 | the tagger's words |
| concept_edges | blacklist regions —causes→ illusion of high correlation status: observational causal language ('leading to'), not an intervention · the relation itself is separately evidenced · strength of the relation: 1 · note: Unfiltered blacklist regions sequester reads, leading to an illusion of high correlation. also in: blacklist regions · illusion of high correlation | concept_edges[5] | 0.95 | the tagger's words |
| concept_edges | identification of these regions and subsequent filtering —causes→ more accurate and stable results across experiments the relation itself is separately evidenced · strength of the relation: 1 · note: The authors explicitly state that identification and filtering lead to the improved result. | concept_edges[6] | 0.95 | the tagger's words |
| concept_edges | removal of blacklists —distinct-from→ typical removal of signals from duplicate reads the relation itself is separately evidenced · strength of the relation: 1 · note: The paper explicitly says blacklist removal differs from duplicate-read removal. | concept_edges[7] | 0.95 | the tagger's words |
| content_tags | A PUBLISHED LIST OF WHAT DOES NOT COUNT, AND APPLYING IT IS MANDATORY: `The removal of the ENCODE blacklist is an essential quality measure when analyzing functional genomics data` | content_tags.findings[0] | 0.95 | the tagger's words |
| content_tags | 582 MILLION OF 2.5 BILLION uniquely aligning reads in the human ENCODE ChIP-seq data fall into blacklisted regions - about 23% of the data, from a small fraction of the mappable genome | content_tags.findings[1] | 0.95 | the tagger's words |
| content_tags | ENCODE uses the proportion of blacklisted reads as a quality control metric, with some experiments reaching 87% | content_tags.findings[2] | 0.95 | the tagger's words |
| content_tags | the class is a PERCENTILE: regions in `the top 1%` of read depth or multi-mapping rate `are considered likely artifacts` | content_tags.findings[3] | 0.95 | the tagger's words |
| content_tags | without blacklist filtering REST, a known repressor, correlates with mostly activating TFs; after filtering the spurious correlation essentially disappears | content_tags.findings[4] | 0.95 | the tagger's words |
| content_tags | and the failure mode is an ILLUSION OF AGREEMENT, not noise: blacklist regions `sequester a large portion of ChIP-seq reads, leading to an illusion of high correlation` | content_tags.findings[5] | 0.95 | the tagger's words |
| content_tags | THE LIST REMOVES ~3% OF HUMAN PROTEIN CODING GENES FROM ANALYSIS: `those studying genes in unmappable regions of the genome will find their data filtered by the blacklist` | content_tags.findings[6] | 0.95 | the tagger's words |
| content_tags | differences between blacklists lead to the hypothesis that artifact regions stem mainly from assembly issues | content_tags.findings[7] | 0.95 | the tagger's words |
| content_tags | and the list is bound to one build: `the blacklists are specific to each genome assembly and a lift over from an old assembly is not meaningful or valid` | content_tags.findings[8] | 0.95 | the tagger's words |
| content_tags | THE RENAMING IS IN THE PAPER: the original list `TERMED the Duke Excluded Regions (DER)`, `now referred to as ENCODE Data Analysis Center (DAC) blacklisted regions` | content_tags.findings[9] | 0.95 | the tagger's words |
| content_tags | and the hand-curated version MISSED AN ENTIRE CRITERION - nuclear mitochondrial DNA segments were `a criterion that was overlooked in the initial manual blacklists` | content_tags.findings[10] | 0.95 | the tagger's words |
| content_tags | problematic regions have generally been left unfiltered because they were not found to affect final signal, but in ChIP-seq they can dramatically bias results | content_tags.findings[11] | 0.95 | the tagger's words |
| content_tags | a comprehensive set of regions in the human, mouse, worm and fly genomes that have anomalous, unstructured, or high signal in next-generation sequencing experiments INDEPENDENT OF CELL LINE OR EXPERIMENT also in: ENCODE blacklist | content_tags.definitions[0] | 0.95 | the tagger's words |
| content_tags | a 1 kb window whose input read depth or multi-mapping read rate is in the TOP 1%, after quantile normalisation across many cell types also in: artifact region | content_tags.definitions[1] | 0.95 | the tagger's words |
| content_tags | per cent - the top of the read-depth and multi-mapping distributions, which is where `artifact` begins | content_tags.quantities[0] | 0.95 | the tagger's words |
| content_tags | per cent of reads falling in blacklisted regions in the worst ENCODE experiments | content_tags.quantities[1] | 0.95 | the tagger's words |
| content_tags | per cent of human protein coding genes that are unmappable and therefore filtered out by the list | content_tags.quantities[2] | 0.95 | the tagger's words |
| content_tags | quality control metric also in: quality control metric | content_tags.statistics[0] | 0.95 | the tagger's words |
Tags that could not be placed in the text:
What was deliberately not tagged, and why, is recorded per paper by the tagger. Where a paper has no such record, this page says so rather than implying that nothing was declined.