palindromes analysis - MCScanX

[KamilSJaron]

• download gene annotation
• reformat, do a check that gene names are matching
• extract proteins sequences using genomes and annotations
• blast all v all
• run MCScanX with two parameter sets (default, palindrome relevant)

[KamilSJaron]

Alternatively, or complementary we could get them also from the MUMmer output. Palindromes on the nucleotide level.

[KamilSJaron]

this analysis is done for all the species with assembly and annotation.

Yet, the outputs need to be parsed.

[KamilSJaron]

I found that a lot of the annotation files do not contain gene annotations.

I am not sure why _genomic.gff.gz file sometimes contains only contigs and sometimes gene annotation. Also some genomes have a separated ftp with annotation at /genomes//GFF/ with a .gff3 file. What makes these genomes special and why there is more than one way to store the annotation? No idea, but I got to curate the files and check that the annotation file actually contain what we need.

Examples: gff = contigs : Lcla1 gff = contigs + genes : Dcor1 separeted gff3 = Obir1

[KamilSJaron]

Species with (some sort of) result: Avag1 Dcor1 Fcan1 Lcla1

For the rest: TODO get protein sequences

[KamilSJaron]

Pvir1 annotation added, maybe incompatible with the NCBI annot.

This is a mess. Most of the genomes have no annotation in NCBI and the NCBI genome has no corresponding scaffold IDs to the annotation that is somewhere on ftp servers.

To finish this task we would have to:

• contact authors / download annotations of the rest of the species
• extract proteins run gffread -g $GENOME_UNZIPED -y $PROTEINS $GFF_UNZIPED, where GENOME_UNZIPED is just the unzipped genome sequence, GFF_UNZIPED is unzipped annotation (I suppose it will work with gtf as well) and PROTEINS is name of the output proteins. If the output file is empty it means that scaffold names in the annotation and genome do not match, or that the annotation does not contain annotation of genes
• reformat annotation file MCScanX is expecting a very peculiar formating of gff file, 4 colums: scf# gene starting_position ending_position. Some examples how to reformat gff3 file with awk are in scripts/prepare_data_for_MCScanX.sh.
• run blast (all proteins to all proteins)
• run MCScanX (both are set up here)

Note that I wrote originally these script for all the genomes, but then I figured out that the scripts are not generally usable, therefore they are sort of "collection of copy-paste commans".

[KamilSJaron]

Missing annotations:

• Anan1 - done

• Mare3 - does not really matter sine we have annotatios for Mare1 and 2, but still. This assembly should be nicer - it's the lattest

• Dpul1 - sexual reference, annotation is sexual annotation

• Aruf1 - annotation does not exist

[jensbast]

Anan1 - Philipp gave me this dropbox link to the GFF https://www.dropbox.com/s/nrx6ccq3d3eaepd/Acrobeloides_nanus_v1.gff3.gz?dl=0

Mare3- the annotation should be on Wormbase within the next 3 weeks he said. He said it is better if we get it from there, because they reformat it and so on.

[KamilSJaron]

MCScanX is done for all but three mentioned above (Anan1 is done).

[KamilSJaron]

TODO: the last one is Mare3

[KamilSJaron]

c89fdd0a82d7eb80e58518f766321c674dadcc5a