Description

The FOX2 CLIP-seq track shows adaptor-trimmed CLIP-seq reads that mapped uniquely to the repeat-masked human genome (hg17). The reads were converted to hg18 coordinates using the UCSC LiftOver tool. Reads on the forward strand are displayed in blue; those on the reverse strand are shown in red.

Methods

Cross-linking immunoprecipitation coupled with high-throughput sequencing (CLIP-seq) of cell type-specific splicing regulator FOX2 (also known as RBM9) was performed in human embryonic stem cells. MosaikAligner was utilized to align the reads to the repeat-masked genome.

Briefly, HUES6 human embryonic stem cells were treated with UV irradiation to stabilize in vivo protein-RNA interactions, followed by antibody-mediated precipitation of specific RNA-protein complexes. SDS-PAGE was then utilized to isolate protein-RNA adducts after RNA trimming with nuclease, 3'RNA linkers were ligated, and nucleotides were 5' end labeled with γ-³²P-ATP. Recovered RNA was ligated to a 5' linker before amplification by RT-PCR. Both linkers were designed to be compatible with Illumina 1G genome analyzer sequencing. Approximately 4 million reads were uniquely mapped to the repeat-masked human genome by MosaikAligner.

To identify CLIP clusters, we performed the following steps: (i) CLIP reads were associated with protein-coding genes as defined by the region from the annotated transcriptional start to the end of each gene locus. (ii) CLIP reads were separated into the categories of sense or antisense to the transcriptional direction of the gene. (iii) Sense CLIP reads were extended by 100 nt in the 5'-to-3' direction. The height of each nucleotide position is the number of reads that overlap that position. (iv) The count distribution of heights is as follows from 1, 2, ...h, ...H-1, H: {n₁, n₂, ...n_h, ...n_H-1, n_H; N = Σn_i (i = 1:H)}. For a particular height, h, the associated probability of observing a height of at least h is P_h = Σn_i(i = h:H) /N. (v) We computed the background frequency after randomly placing the same number of extended reads within the gene for 100 iterations. This controls for the length of the gene and the number of reads. For each iteration, the count distribution and probabilities for the randomly placed reads (P_h,random) was generated as in step (iv). (vi) Our modified FDR for a peak height was computed as FDR(h) = (μ_h + σ_h)/P_h, where μ_h and σ_h is the average and s.d., respectively, of P_h,random across the 100 iterations. For each gene loci, we chose a threshold peak height h* as the smallest height equivalent to FDR(h*) < 0.001. We identified FOX2 binding clusters by grouping nucleotide positions satisfying h > h* and occurred within 50 nt of each other.

For further details of the method used to generate this annotation please refer to Yeo et al. (2009).

Credits

Thanks to Gene Yeo at the University of California, San Diego for providing this annotation. For additional information on FOX2 CLIP-seq reads, please contact geneyeo@ucsd. edu directly.

References

Yeo GW, Coufal NG, Liang YL, Peng GE, Fu XD, Gage FH. An RNA code for the FOX2 splicing regulator revealed by mapping RNA-protein interactions in stem cells. Nat. Struct. Mol. Biol. 2009 Jan 11;16:130-137.