AssemblyHubPT
- Attempted to get UCSC assembly hub working
- Structured assembly hub file structure
- Created proper files
- Work can be seen in the following repo: https://github.com/DanishZahid1/AssemblyHubPT
- Installed fasta to 2 bit converter
conda install -c bioconda ucsc-fatotwobit
conda install -c bioconda/label/cf201901 ucsc-fatotwobit
# Converted fasta to 2bit
faToTwoBit Phaeodactylum_tricornutum.ASM15095v2.dna.toplevel.fa PT.2bit
faToTwoBit: error while loading shared libraries: libssl.so.1.0.0: cannot open shared object file: No such file or directory
[ble: exit 127]
# For some reason I get this error
# Mark suggested using different Conda env (new one)
# That worked
faToTwoBit Phaeodactylum_tricornutum.ASM15095v2.dna.toplevel.fa PT.2bit
<sftp> get PT.2bit
-
Was eventually able to get the assembly hub working
- Tomorrow will add tracks
-
Instructions:
(In danish conda)
- Installed GFF to bed conversion package conda install -c bioconda bedops
(In 2bitConverter conda)
- Installed bedToBigBed program from the binary utilities directory conda install -c bioconda ucsc-bedtobigbed
- To get chrom sizes conda install -c bioconda ucsc-twobitinfo
GFF TO bigBed PROCESS
# Convert GFF to bed
conda activate danish
(danish) dzahid@agrajag:/Volumes/ubdata/dzahid/blastTest/perFeature$ gff2bed < ../../seqFinderTool/phatr3_gene_models_with_href.gff3 > phatr3_gene_models_with_href.bed
# Sort bed
sort -k1,1 -k2,2n phatr3_gene_models_with_href.bed > phatr3_gene_models_with_href_sorted.bed
# remove last 4 columns
cut -d$'\t' -f 7,8,9,10 --complement phatr3_gene_models_with_href_sorted.bed > phatr3_gene_models_with_href_clean.bed
sort -k1,1 -k2,2n phatr3_gene_models_with_href_clean.bed > phatr3_gene_models_with_href_clean_sorted.bed
# Get chrom sizes
conda activate 2bitConverter
twoBitInfo PT.2bit stdout | sort -k2rn > PT.chrom.sizes
# Get bigBed file
bedToBigBed -type=bed6+4 phatr3_gene_models_with_href_sorted.bed PT.chrom.sizes phatr3_gene_models_with_href.bb
# Bizzarley things werent generated a working bb file
# Error message seemed to indicate it was because the bed file had . in it (in columns which it should not)
# Replaced all of these with 0
awk 'BEGIN{OFS=FS="\t"} {gsub("\.","0",$5)}1 {gsub("\.", "Null", $4)}1' phatr3_gene_models_with_href_sorted.bed > PT.bed
# Now got bigBed
# Get bigBed file
bedToBigBed -type=bed6+4 PT.bed PT.chrom.sizes PT.bb
# Fixing that still didn't result in a working bb file
# New error: on using bedToBigBed -type=bed6+4 PT.bed PT.chrom.sizes PT.bb
pass1 - making usageList (90 chroms): 27 millis
Expecting 10 words line 4 of PT.bed got 11
# Tried now fixing the bed file by remove the last 4 columns (extra data that bed dosent usually have)
cut -d$'\t' -f 7,8,9,10 --complement PT.bed > PT_clean.bed
sort -k1,1 -k2,2n PT_clean.bed > PT_clean_sorted.bed
# re-ran
bedToBigBed PT_clean_sorted.bed PT.chrom.sizes PT.bb
# No errors (seemed to work)-
Made a database for the new PT genome (generated 2bit file like described earlier)
-
Finally got the conversion to the bigBed file working
-
Only problem is because files get converted like this gff -> bed -> bigBed.
- When the conversion from gff to bed happens some features don't get named nicely (the 4th column) and so we cant see their names on the annotation track.
- But we can click on them to see their chromosome positions.
- When the conversion from gff to bed happens some features don't get named nicely (the 4th column) and so we cant see their names on the annotation track.
-
But when we just add a custom track by uploading a file through the browser we can use our gff file (no conversions needed) and so we don't run into this issue there.
Instructions:
- Go to https://genome.ucsc.edu/cgi-bin/hgHubConnect?hubId=3613437&hgHub_do_disconnect=on&hgHubConnect.remakeTrackHub=on&hgsid=1440384845_DoCocqhtgEG1XCRVA3ovsM86amaE
- Connect the hub using the following URL:
- To add custom annotation track:
- press add custom tracks
- upload GFF file
