Find orthologs using EggNOG and local HMMER search for multiple transcripts

Question

[Farbod]

Dear Friends Hi ('m not native in English!)

I want to search for orthologs locally using EggNOG in my fish transcriptome samples.

First I have download fiNOG which is specially for fishes from EggNOG,

then built a HMMER database using hmmpress (i.e. cat fiNOG_hmm/*.hmm > fishDB.hmmer), and run hmmpress fishDB.hmmer.

then I intend to run such script : hmmscan --cpu 24 '/home/fiNOG_hmm/fishDB.hmmer' '/home/Transcriptome.fasta'

My question: Can I use my transcriptome assembly .fasta file directly in this script or I must convert (translate) it into protein in advance (which tool is better for this job? Transdecoder ? )?

Thank you

Answer 1

[jhc]

There is now a better resource for functional annotation using eggNOG orthology. Check these links:

Fast genome-wide functional annotation through orthology assignment by eggNOG-mapper

• http://eggnog-mapper.embl.de (online tool)

• http://biorxiv.org/content/early/2016/09/22/076331 (method description and benchmark)

• https://github.com/jhcepas/eggnog-mapper (eggnog-mapper tool. Use the option --translate if using nucleotide seqs)

[Farbod]

Dear jhc, Hi and thank you for your guidance.

It seems that the http://eggnog-mapper.embl.de (online tool) is not suitable for nucleotide seqs, yes?

[jhc]

I just added this option. Nucleotide sequences should now be accepted and automatically translated.

[Farbod]

Thank you jaime,

Yesterday I used the online module and feed it some transcript seqs and used the fiNOG, but unfortunately the result file was empty, I guessed that the online tool did not convert my nucleotide sequence to protein.

[jhc]

no, that option was not enabled yesterday. But, even now, upload protein sequences if possible. translation is automatic and if the CDS sequence is not correctly formatted you will miss annotations.

[Farbod]

Hi again,

Now I have used about 1353 fish transcripts and This time the annotation result contain 144 result (So, I think that works - but I hope more annotation than 144! )

So your suggestion is this that I used the command line with the --transcript option ?

this is the head of my results:

TRINITY_DN246193_c0_g1_i1   7955.ENSDARP00000098851 1.2e-27 85.1    FAM59B          fiNOG[1]    0A18E@biNOG,0DR6J@chorNOG,0II0U@euNOG,0N2VQ@fiNOG,0VC44@meNOG,0XSPW@NOG,12VKE@opiNOG,1CU2B@veNOG    0N2VQ|1.2e-26|93.4  S   family with sequence similarity 59, member B
TRINITY_DN96276_c0_g3_i5    8083.ENSXMAP00000019246 1.2e-177    582.0   ERCC5       map03420    fiNOG[1]    09TAU@biNOG,0DEN5@chorNOG,0N0I4@fiNOG,0V49Z@meNOG,12N9G@opiNOG,1CPC1@veNOG,COG0258@NOG,KOG2520@euNOG    0N0I4|1.3e-226|755.8    L   Excision repair cross-complementing rodent repair deficiency, complementation group 5
TRINITY_DN97247_c0_g1_i10   7955.ENSDARP00000071650 1e-57   184.2   PREB        map04141    fiNOG[1]    09USQ@biNOG,0DG2A@chorNOG,0MX1Q@fiNOG,0V5S6@meNOG,0XRQK@NOG,12SXV@opiNOG,1CQRT@veNOG,KOG0771@euNOG  0MX1Q|1.7e-59|202.2 U   prolactin regulatory element binding

[jhc]

Most probably, many of your nucleotide seqs are not in the correct frame or are not the complete CDS. My suggestion is: translate the set of nucleotide sequences using a dedicated tool, review that the protein products make sense, and upload the protein set to eggnog mapper.

[Farbod]

OK. thank you for your help

Do you have any suggestion for a good "translator tools" ? something that I could input all my transcripts to in and give the protein translation as output (something same as Transdecoder).

~ Best

[urema]

HI,

Which methods would you use for finding orthologs of a particular protein.

Single protein (e.g. P01112) - find orthologs?

I wish to do this offline, with local databases - which resources are best to use here?

Thanks, U.