Frequently asked questions

Frequently asked questions and their answers about the refTSS resource

Q1. Is there dataset of transcription start sites collected with only noncording RNA regions?

• A1. Currently, there are no data that collects only the transcription start sites of non-coding RNAs. It is possible to extract noncording RNAs by concatenating the GENCODE annotation file and the following files:
  http://reftss.clst.riken.jp/datafiles/3.0/human/refTSS_v3.0_human_annotation.txt.gz
  http://reftss.clst.riken.jp/datafiles/3.0/mouse/refTSS_v3.0_mouse_annotation.txt.gz

Q2. Why the range of some TSSs are so much greater than others? For most TSSs the range is in the region of 8-24 bp, while for others the range is over 100 bp.

• A2. The range of TSSs is in the region of 8-24 bp, while for others the range can be over 100 bp. We computationally identify TSS regions based on the mapping results with 5'-end sequences.

Q3. most TSSs with an associated gene symbol are within ~20 bp of the 5' end of a known transcript, but how much confidence can we have that TSSs that are much further away than this truly represent a TSS for the associated gene?

• A3. We search for potential associated genes based on our annotation pipeline. In this pipeline, we searched for annotated TSS within 500 bp, and choose the nearest genes as associated ones. In this case, we consider the TSS is an unannotated (undetected) 1st exon. Of course, since it is a prediction, it can be incorrect. We have no clear guidelines about the confidence, but the distance can be a quasi-measure of the strength of the prediction.

Q4. Could you please provide with more information about the GRCh38.p10 file?

• A4. Usually, even if the patch number is different, the coordination is not changed. You can use the current TSSs coordinates provided in refTSS regardless of the patch number.

Q5. How to go about making a GTF from the provided refTSS BED files?

• A5. There are many ways in which you can convert BED file formats to GTF. For example, a command that uses UCSC tools (bedToGenePred and genePredToGtf) like this:

  >	bedToGenePred in.bed /dev/stdout | genePredToGtf file /dev/stdin out.gtf

  
  If you are using Galaxy, it provides solution for converting BED to GTF.

Q6. When looking at the TSS sites on your refTSS track in the UCSC Genome Browser, some appear blue and others red. Why is this difference? On the other hand, the number of white arrow heads within the TSS sites is indicative of what?

• A6. In refTSS’ UCSC configuration of the bigBed 6 tracks, we set the option:

  colorByStrand 255,0,0 0,0,255

  This setting enables to show alignments on the reverse strand colored in dark red, whereas alignments on the forward strand are colored in dark blue. Again, in refTSS UCSC configuration of the bigBed 6 tracks, we set the option:

  exonArrows on
  exonArrowsDense on

  The above options enable us to show the Exon arrows in white heads within the TSS.

Q7. Does the refTSS project have bigWig file format to verify strengths of TSS peak?

• A7. In the current release of refTSS we only provide BED and TEXT formats, but not bigWig files. We are planning to offer this in a future release. However, the BED files from refTSS can be used to generate bigwig file format using several UCSC tools (http://hgdownload.soe.ucsc.edu/admin/exe/) that supports this kind of conversion.

Q8. Could you point me to the direction of the bed files containing TSS peak-score info, or the raw data?

• A8. In refTSS we only integrate TSS regions, but not the TSS tag counts (Score), this is due to the fact that different data sources use different sequence specifications. Instead, we recommend you take the expression tables from the FANTOM5 web site (https://fantom.gsc.riken.jp/5/datafiles/reprocessed/mm10_latest/extra/CAGE_peaks_expression/) and then retrieve the count data.

Q9. Is there a possibility to determine the location of the translational start site and the premature stop codon according to the refTSS sequence?

• A9. In the current version of refTSS, you can’t determine the location of the start and stop codons, but you can predict them using the transcript and protein annotation information provided by refTSS at http://reftss.clst.riken.jp/datafiles/current/mouse/gene_annotation/

Q10. Is it possible to get the cell or tissue information for the TSS regions in the database?

• A10. In the current release of refTSS we are not providing sample information (i.e. cell or tissue etc.). If you need such info, you will need to look at the source of TSS from here http://reftss.clst.riken.jp/datafiles/current/human/sources/