You are currently viewing a new version of our website. To view the old version click .
Cells
  • Correction
  • Open Access

5 August 2019

Correction: Nikitin, D., et al. Retroelement—Linked Transcription Factor Binding Patterns Point to Quickly Developing Molecular Pathways in Human Evolution. Cells 2019, 8, 130

,
,
,
,
,
,
,
,
and
1
I.M. Sechenov First Moscow State Medical University, 119992 Moscow, Russia
2
Omicsway Corp., Walnut, CA 91798, USA
3
D. Rogachev Federal Research Center of Pediatric Hematology, Oncology and Immunology, 117198 Moscow, Russia
4
Shemyakin-Ovchinnikov Institute of Bioorganic Chemistry, 117198 Moscow, Russia
In the article ‘Retroelement—Linked Transcription Factor Binding Patterns Point to Quickly Developing Molecular Pathways in Human Evolution,’ a number of transcription factor binding sites (TFBS) mapped on all retroelement classes were incorrectly calculated as sum of TFBS numbers separately mapped on LINEs, SINEs and LTR retrotransposons/endogenous retroviruses (LR/ERVs). This made the proportion of RE-linked TFBS 72.1% (incorrect) instead of 55.5% (correct) of all mapped TFBS. Similarly, the proportion of RE-linked TFBS was incorrectly calculated as 61.4% (incorrect) instead of 47% (correct) of all TFBS mapped in 10 kb gene transcriptional start neighborhood. Corrections were made in the text and Table 1 and Table 2.
Table 1. Overall transcription factor binding sites (TFBS) statistics.
Table 2. Gene neighborhood-linked TFBS statistics.
For example, in Table 1, the total TFBS numbers mapped on SINEs, LINEs and LTRs in all 13 cell lines tested were 86,583,508, 78,331,554 and 3,009,962, respectively. The total number of mapped TFBS in 13 cell lines was 277,187,723. The number of TFBS mapped on all retroelements classes (LINEs, SINEs and LR/ERVs) was incorrectly calculated as the simple arithmetic sum of these three numbers, i.e., 199,925,024, or 72.1% of all mapped TFBS. Similarly, a proportion of gene neighboring TFBS was calculated as 61.4% or all TFBS. This method of calculation was incorrect because it did not take into account that one TFBS could simultaneously overlap with several individual LINE, SINE or LR/ERV elements.
Here, we made corrections by directly mapping TFBS on all retroelement classes. The corrected TFBS numbers are shown on Table 1.
Table 1 should read as follows:
Similarly, we re-calculated corrected the number and percentage of TFBS mapped on all retroelement classes in 10 kb gene neighborhoods (Table 2).
Table 2 should read as follows:
The passage on the page 7 between the Table 1 and Table 2 should read as follows:
“In total, 277,187,723 TFBS hits could be mapped on the human genome for all these cell lines. Of them, 55.5% overlapped with the RE sequences, thus confirming that REs serve as the major source of TFBS in human cells. Considering previous reports, this proportion may seem high [60]. However, in our analysis, all multimapped TFBS reads were filtered out according to the standard ENCODE ChIP-seq mapping and filtering pipeline [42], so the results represented uniquely mapped TFBS reads. To confirm this TFBS proportion, we also validated the method used by parallel mapping of all human RE sequences extracted from USCS Genome Browser [55] on the human genome. In a good agreement with previously published data [], REs mapped by the same approach occupied ~45% of human DNA, genome assembly hg19. We therefore found no technological drawbacks here and suggest that the proportion of 55.5% RE-linked TFBS is correct at least for the ENCODE primary cell culture datasets used. However, this proportion was somewhat lower for the TFBS mapped in a 10 kb neighborhood of gene transcriptional start sites (TSS): Among 43,042,026 totally mapped TFBS, only 47% overlapped with REs. This overall trend was representative for all cell lines under investigation (Table 2). The TSS-proximal TFBS hits were unevenly distributed among the major classes of REs: ~30% were attributed to SINEs; ~17%–to LINEs, and ~7%–to LTR retrotransposons and endogenous retroviruses. For the total fraction of RE-linked TFBS hits (not only gene-proximal), these proportions were, respectively, 28%, 26% and 12%.”
The beginning of the last paragraph in “Background” section should read as follows:
“To distinguish cell type-specific and general evolutionary features, in this study, we investigated the distributions of RE-linked hits for all 13 human cell lines TFBS-profiled for 563 DNA-binding proteins during the ENCODE project [40], representing eight different tissues/organs from the different individuals (Supplementary file 2, Figure S2A). We found that 55.5% of totally mapped TFBS overlapped with the RE sequences, thus confirming RE’s status as the major source of TFBS for human cells.”

Conflicts of Interest

The authors declare no conflict of interest.

References

  1. Nikitin, D.; Garazha, A.; Sorokin, M.; Penzar, D.; Tkachev, V.; Markov, A.; Gaifullin, N.; Borger, P.; Poltorak, A.; Buzdin, A.A. Retroelement-Linked Transcription Factor Binding Patterns Point to Quickly Developing Molecular Pathways in Human Evolution. Cells 2019, 8, 130. [Google Scholar] [CrossRef] [PubMed]

Article Metrics

Citations

Article Access Statistics

Multiple requests from the same IP address are counted as one view.