rRNA track is short fragments?
277 views
Skip to first unread message
J. C. Szamosi
Mar 30, 2016, 1:57:05 PM3/30/16
to gen...@soe.ucsc.edu
Hello all,
I downloaded the human rRNA track following the instructions here:
https://groups.google.com/a/soe.ucsc.edu/forum/#!searchin/genome/rrna/genome/0X06cAZgHjU/YekSLOG7CQAJ
And did the same for mouse following similar instructions. In both cases, the gtf files that I downloaded are annotating a large number of very small sequences (50 - 150bp in length). I find this very surprising. Are the whole rRNA regions not annotated as full, long sequences? The short sequences in the gtf file aren't even contiguous, so it doesn't even cover a long rRNA sequence as I would expect.I downloaded the human rRNA track following the instructions here:
https://groups.google.com/a/soe.ucsc.edu/forum/#!searchin/genome/rrna/genome/0X06cAZgHjU/YekSLOG7CQAJ
--
J. C. Szamosi, M.Sc.
Bioinformatician
Farncombe Metagenomics Facility
McMaster University
Farncombe Metagenomics Facility
McMaster University
Matthew Speir
Mar 31, 2016, 1:52:59 PM3/31/16
to J. C. Szamosi, gen...@soe.ucsc.edu
Hi Jake,
Thank you for your question about obtaining coordinates for rRNA genes.
Many of the results generated by Table Browser query in the answer to the previous mailing list question are 5S rRNA pseudogenes. You can see these types of details by examining one of these annotations in the Genome Browser, clicking into the GENCODE details page for the item and then clicking through to the Ensembl website, where next to "Description" you should see some text like " RNA, 5S ribosomal pseudogene 40". For example http://www.ensembl.org/Homo_sapiens/Transcript/Summary?db=core;g=ENSG00000252956;r=1:9437669-9437778;t=ENST00000517147.
I believe that many of these pseudogenes can be removed from the output by improving the Table Browser filters applied to the results. In step 6 of those outlined in the answer here: https://groups.google.com/a/soe.ucsc.edu/d/msg/genome/0X06cAZgHjU/Kn0WPsAiCgAJ, you can add the following filter under the hg38.wgEncodeGencodeAttrsV22 based filters' to those already described:
Additionally, this list of rRNA genes may not be an exhaustive list of all rRNA genes throughout the genome and only represents those rRNA genes that GENCODE has annotated. From a quick scan of the gene names output using the instructions in the previous mailing list answer plus the additional geneName filter I described, it looks like GENCODE only includes 5S and 5.8S rRNA genes in their annotations. According to the "Ribosomal RNA" Wikipedia page, https://en.wikipedia.org/wiki/Ribosomal_RNA#Prokaryotes_vs._eukaryotes, it looks like the 5S and 5.8S genes are expected to be between 100-156 bp in length.
I hope this is helpful. If you have any further questions, please reply to gen...@soe.ucsc.edu. All messages sent to that address are archived on a publicly-accessible Google Groups forum. If your question includes sensitive data, you may send it instead to genom...@soe.ucsc.edu.
Matthew Speir
UCSC Genome Bioinformatics Group
Thank you for your question about obtaining coordinates for rRNA genes.
Many of the results generated by Table Browser query in the answer to the previous mailing list question are 5S rRNA pseudogenes. You can see these types of details by examining one of these annotations in the Genome Browser, clicking into the GENCODE details page for the item and then clicking through to the Ensembl website, where next to "Description" you should see some text like " RNA, 5S ribosomal pseudogene 40". For example http://www.ensembl.org/Homo_sapiens/Transcript/Summary?db=core;g=ENSG00000252956;r=1:9437669-9437778;t=ENST00000517147.
I believe that many of these pseudogenes can be removed from the output by improving the Table Browser filters applied to the results. In step 6 of those outlined in the answer here: https://groups.google.com/a/soe.ucsc.edu/d/msg/genome/0X06cAZgHjU/Kn0WPsAiCgAJ, you can add the following filter under the hg38.wgEncodeGencodeAttrsV22 based filters' to those already described:
geneName doesn't match *5SP*
Additionally, this list of rRNA genes may not be an exhaustive list of all rRNA genes throughout the genome and only represents those rRNA genes that GENCODE has annotated. From a quick scan of the gene names output using the instructions in the previous mailing list answer plus the additional geneName filter I described, it looks like GENCODE only includes 5S and 5.8S rRNA genes in their annotations. According to the "Ribosomal RNA" Wikipedia page, https://en.wikipedia.org/wiki/Ribosomal_RNA#Prokaryotes_vs._eukaryotes, it looks like the 5S and 5.8S genes are expected to be between 100-156 bp in length.
I hope this is helpful. If you have any further questions, please reply to gen...@soe.ucsc.edu. All messages sent to that address are archived on a publicly-accessible Google Groups forum. If your question includes sensitive data, you may send it instead to genom...@soe.ucsc.edu.
Matthew Speir
UCSC Genome Bioinformatics Group
--
Matthew Speir
Apr 5, 2016, 6:13:01 PM4/5/16
to J. C. Szamosi, gen...@soe.ucsc.edu
Hi Jake,
Thank you for the suggestion, I will make sure it's something that's taken into consideration for future responses concerning getting rRNA gene coordinates from the GENCODE track in the UCSC Genome Browser.
As for a track that may have a complete set of rRNA annotations, I'm not sure that we have one our public site. I was going to suggest the RefSeq Genes track, but even that only appears to contain a subset of the possible rRNA gene annotations. According to the GenBank description of the 28S rRNA, https://www.ncbi.nlm.nih.gov/nuccore/NR_003287.2?report=genbank, the regions containing the the 45S rRNA precursor for the 18S, 5.8S and 28S rRNA should be found on chromosomes 13, 14, 15, 21 and 22. However, our RefSeq Genes track only contains annotations for these three rRNA genes on chr21, an unplaced chr22 scaffold, and an unlocalized and unplaced scaffold:
Even though this output contains a few more annotations for the larger rRNA genes, I doubt that this would be a complete annotation of all the rRNA genes in the genome. This is because, according to Wikipedia, each cluster on chromosomes 13, 14, 15, 21, and 22 contains 30-40 repeats of the 45S rRNA precursor gene whereas the above list only contains one 45S annotation per chromosome and only a few 5.8S, 18S, and 28S annotations outside of these. That's not not even counting those rRNA annotations missing from chr13, chr14, and chr15. You can try asking this question of other, more general online biology help forums including Biostars or SeqAnswers to see if others have recommendations for finding a full set of all rRNA genes in the genome or how to best mask these rRNA genes from your RNA-seq analysis.
Thank you for the suggestion, I will make sure it's something that's taken into consideration for future responses concerning getting rRNA gene coordinates from the GENCODE track in the UCSC Genome Browser.
As for a track that may have a complete set of rRNA annotations, I'm not sure that we have one our public site. I was going to suggest the RefSeq Genes track, but even that only appears to contain a subset of the possible rRNA gene annotations. According to the GenBank description of the 28S rRNA, https://www.ncbi.nlm.nih.gov/nuccore/NR_003287.2?report=genbank, the regions containing the the 45S rRNA precursor for the 18S, 5.8S and 28S rRNA should be found on chromosomes 13, 14, 15, 21 and 22. However, our RefSeq Genes track only contains annotations for these three rRNA genes on chr21, an unplaced chr22 scaffold, and an unlocalized and unplaced scaffold:
In addition to these annotations, there are a number of 5S rRNAs. The RefSeq Genes track is based on aligning RNAs from RefSeq to the genome and then selecting those best alignments. In cases where there are two best alignments, both are kept. I know that GENCODE annotated a number of 5S pseudogenes, so it's possible that some of the multiple alignments for a particular 5S gene could overlap with some of the GENCODE 5S pseudogenes, though I haven't investigated this. You can read more about how the RefSeq Genes track is constructed on the track description page here: http://genome.ucsc.edu/cgi-bin/hgTrackUi?db=hg38&g=refGene. If you're interested, I can give some instructions on how to extract rRNA annotations from the RefSeq Genes track.NR_003287 chr21 + 8213887 8401980 rRNA RNA28S5 NR_003285 chr21 + 8212571 8212727 rRNA RNA5-8S5 NR_003285 chr21 + 8256780 8256936 rRNA RNA5-8S5 NR_003286 chr21 + 8209630 8211499 rRNA RNA18S5 NR_003285 chr21 + 8395606 8395762 rRNA RNA5-8S5 NR_003285 chr21 + 8439822 8439978 rRNA RNA5-8S5 NR_046235 chr21 + 8433221 8446572 rRNA RNA45S5 NR_003287 chr21 + 8441145 8446211 rRNA RNA28S5 NR_003286 chr21 + 8436875 8438744 rRNA RNA18S5 NR_003286 chr21 + 8392665 8394534 rRNA RNA18S5 NR_003287 chrUn_GL000220v1 + 113347 118417 rRNA RNA28S5 NR_046235 chrUn_GL000220v1 + 105423 118780 rRNA RNA45S5 NR_003285 chrUn_GL000220v1 + 112024 112180 rRNA RNA5-8S5 NR_003286 chrUn_GL000220v1 + 109077 110946 rRNA RNA18S5 NR_003285 chrUn_GL000220v1 + 155996 156152 rRNA RNA5-8S5 NR_003286 chrUn_GL000220v1 + 153049 154918 rRNA RNA18S5 NR_003287 chr22_KI270733v1_random + 130203 135280 rRNA RNA28S5 NR_046235 chr22_KI270733v1_random + 122272 135645 rRNA RNA45S5 NR_003285 chr22_KI270733v1_random + 128876 129032 rRNA RNA5-8S5 NR_003286 chr22_KI270733v1_random + 125930 127799 rRNA RNA18S5 NR_003285 chr22_KI270733v1_random + 173955 174111 rRNA RNA5-8S5 NR_003286 chr22_KI270733v1_random + 171011 172880 rRNA RNA18S5
Even though this output contains a few more annotations for the larger rRNA genes, I doubt that this would be a complete annotation of all the rRNA genes in the genome. This is because, according to Wikipedia, each cluster on chromosomes 13, 14, 15, 21, and 22 contains 30-40 repeats of the 45S rRNA precursor gene whereas the above list only contains one 45S annotation per chromosome and only a few 5.8S, 18S, and 28S annotations outside of these. That's not not even counting those rRNA annotations missing from chr13, chr14, and chr15. You can try asking this question of other, more general online biology help forums including Biostars or SeqAnswers to see if others have recommendations for finding a full set of all rRNA genes in the genome or how to best mask these rRNA genes from your RNA-seq analysis.
I hope this is helpful. If you have any further questions, please reply to gen...@soe.ucsc.edu. All messages sent to that address are archived on a publicly-accessible Google Groups forum. If your question includes sensitive data, you may send it instead to genom...@soe.ucsc.edu.
Matthew Speir
UCSC Genome Bioinformatics Group
On 4/4/16 7:28 AM, J. C. Szamosi wrote:
JakeThanks!Thanks for your help, Matthew.It seems that if only the 5S and 5.8S subunits are annotated, that this implies people should be aware of that fact when using that track as a mask file for RNASeq experiments. Do you know of a track that annotates all the rRNA in human or mouse genomes? And if not, it might be worth mentioning this in future posts providing the previously linked instructions, so investigators aren't led astray.
J. C. Szamosi
Apr 6, 2016, 10:11:15 AM4/6/16
to gen...@soe.ucsc.edu
Thanks so much for your help, Matthew!
Reply all
Reply to author
Forward
0 new messages