Tracks in Custom genome

[YeHW]

Hi, igv team.

I've got human genome fasta file (human 1kg_v37) and want to setup a custom genome.json for that to use with igv. I've checked wiki and b37_1kg.json shipped with igv and had some questions about the fields in the json file.

1. In b37_1kg.json, "cytobandURL": "https://s3.amazonaws.com/igv.org.genomes/1kg_v37/b37_cytoband.txt", what is the origin of this cytoband file? I found that the cytoband for GRCh38 is from UCSC. Is it the same case for 1kg_v37?
2. In b37_1kg.json, how is "url": "https://s3.amazonaws.com/igv.org.genomes/hg19/ncbiRefSeq.sorted.txt.gz" in the Refseq Genes track sorted and tabixed? I found there's only a unsorted ncbiRefSeq.txt.gz in ucsc's ftp site.
3. In the above Refseq Genes track, there is only 1 transcript (NM_002944.2) for ROS1 gene, but in the ncbi GDV there are 3 transcripts (NM_002944.3, NM_001378902.1, NM_001378891.1) for ROS1 gene (because it's using the latest Annotation Release 105.20220307). How can I build a new Refseq Genes track to be used with igv using the latest Annotation Release from Refseq (105.20220307 as of writing)?

Thank you!

[maximilianh]

1. In b37_1kg.json <https://s3.amazonaws.com/igv.org.genomes/1kg_v37/b37_1kg.json>, "cytobandURL": "https://s3.amazonaws.com/igv.org.genomes/1kg_v37/b37_cytoband.txt", what is the origin of this cytoband file? I found that the cytoband for GRCh38 is from UCSC. Is it the same case for 1kg_v37? I am not with IGV, but I bet the cytoband file also comes from UCSC. 1. In b37_1kg.json <https://s3.amazonaws.com/igv.org.genomes/1kg_v37/b37_1kg.json>, how is "url": " https://s3.amazonaws.com/igv.org.genomes/hg19/ncbiRefSeq.sorted.txt.gz" in the Refseq Genes track sorted and tabixed? I found there's only a unsorted ncbiRefSeq.txt.gz in ucsc's ftp <https://hgdownload.cse.ucsc.edu/goldenpath/hg19/database/> site. 1. In the above Refseq Genes track, there is only 1 transcript (NM_002944.2) for ROS1 gene, but in the ncbi GDV <https://www.ncbi.nlm.nih.gov/genome/gdv/browser/genome/?id=GCF_000001405.25> there are 3 transcripts for ROS1 gene (because it's using the latest Annotation Release 105.20220307). How can I build a new *Refseq Genes* track to be used with igv using the latest Annotation Release from Refseq ( 105.20220307 <https://ftp.ncbi.nlm.nih.gov/refseq/H_sapiens/annotation/annotation_releases/105.20220307/GCF_000001405.25_GRCh37.p13/> as of writing)? UCSC also uses 105.20220307 on hg19 and there is only one location

for NM_002944.3 chr6 - 117608515 - 117747105. The GDV also has only one location for NM_002944.3. On hg38, and T2T, only one location in GDV. On the UCSC search, for hg38, there are three locations, but on three different tracks (UCSC's mapping, NCBI's mapping and Gencode)

…

Thank you! — Reply to this email directly, view it on GitHub <#1306>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AACL4TKZFO6DUZVHP53KXVDW6KSDPANCNFSM6AAAAAAWKIGFMU> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[jrobinso]

Thanks @maximilianh , you are correct, the cytoband files all come from UCSC.

Tabix indexing is optional, and not really necessary for tracks < 20MB in total size. I would suggest skipping this, but if you have a need for it documentation is here: https://www.htslib.org/doc/tabix.html.

I assume you have reviewed the documentation (https://github.com/igvteam/igv.js/wiki) that describes genomes.json and other files.

[YeHW]

1. In b37_1kg.json https://s3.amazonaws.com/igv.org.genomes/1kg_v37/b37_1kg.json, "cytobandURL": "https://s3.amazonaws.com/igv.org.genomes/1kg_v37/b37_cytoband.txt", what is the origin of this cytoband file? I found that the cytoband for GRCh38 is from UCSC. Is it the same case for 1kg_v37? I am not with IGV, but I bet the cytoband file also comes from UCSC. 1. In b37_1kg.json https://s3.amazonaws.com/igv.org.genomes/1kg_v37/b37_1kg.json, how is "url": " https://s3.amazonaws.com/igv.org.genomes/hg19/ncbiRefSeq.sorted.txt.gz" in the Refseq Genes track sorted and tabixed? I found there's only a unsorted ncbiRefSeq.txt.gz in ucsc's ftp https://hgdownload.cse.ucsc.edu/goldenpath/hg19/database/ site. 1. In the above Refseq Genes track, there is only 1 transcript (NM_002944.2) for ROS1 gene, but in the ncbi GDV https://www.ncbi.nlm.nih.gov/genome/gdv/browser/genome/?id=GCF_000001405.25 there are 3 transcripts for ROS1 gene (because it's using the latest Annotation Release 105.20220307). How can I build a new Refseq Genes track to be used with igv using the latest Annotation Release from Refseq ( 105.20220307 https://ftp.ncbi.nlm.nih.gov/refseq/H_sapiens/annotation/annotation_releases/105.20220307/GCF_000001405.25_GRCh37.p13/ as of writing)? UCSC also uses 105.20220307 on hg19 and there is only one location
   for NM_002944.3 chr6 - 117608515 - 117747105. The GDV also has only one location for NM_002944.3. On hg38, and T2T, only one location in GDV. On the UCSC search, for hg38, there are three locations, but on three different tracks (UCSC's mapping, NCBI's mapping and Gencode)
   …
   Thank you! — Reply to this email directly, view it on GitHub <Tracks in Custom genome #1306>, or unsubscribe https://github.com/notifications/unsubscribe-auth/AACL4TKZFO6DUZVHP53KXVDW6KSDPANCNFSM6AAAAAAWKIGFMU . You are receiving this because you are subscribed to this thread.Message ID: @.***>

Thanks @maximilianh and @jrobinso.

Just want to make sure about one thing:
I spot three NM_ transcripts for ROS1 gene in GDV:

I guess it's because there are three NM_ transcripts in the RefSeq release 105.20220307. I downloaded GCF_000001405.25_GRCh37.p13_genomic.gtf.gz from the above ftp site, and checked:

curl -s 'https://ftp.ncbi.nlm.nih.gov/refseq/H_sapiens/annotation/annotation_releases/105.20220307/GCF_000001405.25_GRCh37.p13/GCF_000001405.25_GRCh37.p13_genomic.gtf.gz' \
  | zcat GCF_000001405.25_GRCh37.p13_genomic.gtf.gz \
  | rg '(gene_id "ROS1"; transcript_id "NM_.*?")' -or '$1' \
  | sort | uniq

# output
## gene_id "ROS1"; transcript_id "NM_001378891.1"
## gene_id "ROS1"; transcript_id "NM_001378902.1"
## gene_id "ROS1"; transcript_id "NM_002944.3"

If the above speculation is correct, I want to build a new igv track based on RefSeq release 105.20220307. To ahieve that, I need to build a file like ncbiRefSeq.sorted.txt.gz. Could you please help me with this?

Thank you!

[maximilianh]

"I spot three NM_ transcripts for ROS1 gene in GDV" - OK, now this is entirely different. You originally said that there were three locations for the transcript NM_002944.3. Yes, I think your analysis is correct. The updated file ncbiRefSeq for 105.20220307 is here https://hgdownload.soe.ucsc.edu/goldenPath/hg38/database/ncbiRefSeq.txt.gz and as Jim mentioned, you can sort it but that's not required. You should be able to load the file as-is, if I understood him correctly (but he may be able to confirm that. Jim: we're updating these files on a regular schedule now, automatically. Maybe we can figure out a way to let you know when to update these? We can send email or ping a URL when we update. We also keep the previous versions, so you could in theory tag them with the release name and offer a version history.

…

On Thu, Mar 30, 2023 at 11:59 AM Hongwei Ye ***@***.***> wrote: 1. In b37_1kg.json https://s3.amazonaws.com/igv.org.genomes/1kg_v37/b37_1kg.json, "cytobandURL": " https://s3.amazonaws.com/igv.org.genomes/1kg_v37/b37_cytoband.txt", what is the origin of this cytoband file? I found that the cytoband for GRCh38 is from UCSC. Is it the same case for 1kg_v37? I am not with IGV, but I bet the cytoband file also comes from UCSC. 1. In b37_1kg.json https://s3.amazonaws.com/igv.org.genomes/1kg_v37/b37_1kg.json, how is "url": " https://s3.amazonaws.com/igv.org.genomes/hg19/ncbiRefSeq.sorted.txt.gz" in the Refseq Genes track sorted and tabixed? I found there's only a unsorted ncbiRefSeq.txt.gz in ucsc's ftp https://hgdownload.cse.ucsc.edu/goldenpath/hg19/database/ site. 1. In the above Refseq Genes track, there is only 1 transcript (NM_002944.2) for ROS1 gene, but in the ncbi GDV https://www.ncbi.nlm.nih.gov/genome/gdv/browser/genome/?id=GCF_000001405.25 there are 3 transcripts for ROS1 gene (because it's using the latest Annotation Release 105.20220307). How can I build a new *Refseq Genes* track to be used with igv using the latest Annotation Release from Refseq ( 105.20220307 https://ftp.ncbi.nlm.nih.gov/refseq/H_sapiens/annotation/annotation_releases/105.20220307/GCF_000001405.25_GRCh37.p13/ as of writing)? UCSC also uses 105.20220307 on hg19 and there is only one location for NM_002944.3 chr6 - 117608515 - 117747105. The GDV also has only one location for NM_002944.3. On hg38, and T2T, only one location in GDV. On the UCSC search, for hg38, there are three locations, but on three different tracks (UCSC's mapping, NCBI's mapping and Gencode) … <#m_-8401278165509756520_> Thank you! — Reply to this email directly, view it on GitHub <#1306 <#1306>>, or unsubscribe https://github.com/notifications/unsubscribe-auth/AACL4TKZFO6DUZVHP53KXVDW6KSDPANCNFSM6AAAAAAWKIGFMU . You are receiving this because you are subscribed to this thread.Message ID: *@*.***> Thanks @maximilianh <https://github.com/maximilianh> and @jrobinso <https://github.com/jrobinso>. Just want to make sure about one thing: I spot three NM_ transcripts for ROS1 gene in GDV <https://www.ncbi.nlm.nih.gov/genome/gdv/browser/genome/?id=GCF_000001405.25> : [image: image] <https://user-images.githubusercontent.com/43214065/228794780-46125249-ca93-49c4-bae3-b2a0f8f46ac4.png> I guess it's because there are three NM_ transcripts in the RefSeq release 105.20220307 <https://ftp.ncbi.nlm.nih.gov/refseq/H_sapiens/annotation/annotation_releases/105.20220307/GCF_000001405.25_GRCh37.p13/>. I downloaded GCF_000001405.25_GRCh37.p13_genomic.gtf.gz <https://ftp.ncbi.nlm.nih.gov/refseq/H_sapiens/annotation/annotation_releases/105.20220307/GCF_000001405.25_GRCh37.p13/GCF_000001405.25_GRCh37.p13_genomic.gtf.gz> from the above ftp site, and checked: curl -s 'https://ftp.ncbi.nlm.nih.gov/refseq/H_sapiens/annotation/annotation_releases/105.20220307/GCF_000001405.25_GRCh37.p13/GCF_000001405.25_GRCh37.p13_genomic.gtf.gz' \ | zcat GCF_000001405.25_GRCh37.p13_genomic.gtf.gz \ | rg '(gene_id "ROS1"; transcript_id "NM_.*?")' -or '$1' \ | sort | uniq # output## gene_id "ROS1"; transcript_id "NM_001378891.1"## gene_id "ROS1"; transcript_id "NM_001378902.1"## gene_id "ROS1"; transcript_id "NM_002944.3" If the above speculation is correct, I want to build a new igv track based on RefSeq release 105.20220307 <https://ftp.ncbi.nlm.nih.gov/refseq/H_sapiens/annotation/annotation_releases/105.20220307/GCF_000001405.25_GRCh37.p13/>. To ahieve that, I need to build a file like *ncbiRefSeq.sorted.txt.gz*. Could you please help me with this? Thank you! — Reply to this email directly, view it on GitHub <#1306 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AACL4TOCOIT6RNSITZRE4Q3W6VKR3ANCNFSM6AAAAAAWKIGFMU> . You are receiving this because you were mentioned.Message ID: ***@***.***>

[jrobinso]

@maximilianh Thanks for the info. Actually I was considering just referencing those URLs directly (e.g. https://hgdownload.soe.ucsc.edu/goldenPath/hg38/database/ncbiRefSeq.txt.gz). The only thing I do additionally is tabix index them, but really they aren't that large and the benefit is marginal.

[jrobinso]

@YeHW You do not need to build a file like ncbiRefSeq.sorted.txt.gz. See the user documentation, all common annotation formats are supported, including gff3.

[maximilianh]

Yes, that would be perfect. We have never, not once, changed these URLs so far.

…

On Thu, Mar 30, 2023 at 5:37 PM Jim Robinson ***@***.***> wrote: @YeHW <https://github.com/YeHW> You do not need to build a file like ncbiRefSeq.sorted.txt.gz. See the user documentation, all common annotation formats are supported, including gff3. — Reply to this email directly, view it on GitHub <#1306 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AACL4TPRG7WWYRBAVJVHYATW6WSDHANCNFSM6AAAAAAWKIGFMU> . You are receiving this because you were mentioned.Message ID: ***@***.***>

[jrobinso]

This is done. @YeHW if you update your genome by selecting "Genomes > Select Hosted Genome" from the menu you will get updated annotation. The assembly you are asking about is in the updated menu as follows

[YeHW]

@maximilianh @jrobinso Thank you! I can see the updated annotation.