3
$\begingroup$

I want a process by which I can identify genes within plant genomes; i.e. a process that works for all plants, not just for a single species.

I'm wondering, if, by bioinformatic means only, it is possible to "extract" genes and, after that proteins and natural compounds from plant genome data that is available on NCBI (please see https://www.ncbi.nlm.nih.gov/datasets/genome/GCA_029618835.1/ as an example)? Example: just by processing the data from the link above we can get to Flavoxanthin or any other substance.

What other useful information from a chemical compound perspective can be found from digging into the genome itself?

Any hints will be greatly appreciated.

Thank you!

$\endgroup$
0

1 Answer 1

2
+50
$\begingroup$

I believe I understand the project, to obtain a biochemical/metabolic profile of a large number of poorly annotated plant genomes. There isn't a perfect solution, like Prokka for bacteria. It is clearly an area subject to recent development and interest.

I would recommend using Gene Ontology Meta Annotator for Plants (GOMAP) which is described in "Plant Methods" journal.

There are limitations:

  • KEGG is not used (below);
  • its a singularity container - and thats its only distribution.

Learning Singularlity You will need to learn running the container through Singularity and that will be limiting if you are not familiar with containers. However, this is the best of the available "out of the box" solutions I can see in plants. A Singularity HPC is a good call here, tricky if you don't have access to one. I know Docker, Singularity is similar but different. The information on Singularity is here. I would start reading from "Executing Commands"

singularity pull docker://godlovedc/lolcow # get the container 
singularity exec lolcow_latest.sif cowsay moo # run it

(the cow is a famous teaching tool for containers)

KEGG One approach that can be used is to perform COG (NCBI/Genbank code) and KEGG no. identification (e.g. eggNOG). Tools such as eggNOG are setup for metagenomics data however. The COG/KEGG no. output is only the starting point, interrogating each of these to assign gene function and metabolic pathways will require a KEGG database and custom code. Unless someone has published, e.g. KEGG database in SQL.

In my experience the Kyoto Encyclopaedia of Genes and Genomes (KEGG) is the best singular functional database available, thats why I see GOMAPs exclusion of this as a limitation. COG doesn't provide a particularly useful information functional description IMO. eggNOG, or any other method to identify COGs and KEGG numbers, interrogated against the KEGG database works for plants very well. I know this because I'm not focused on plants but often recover plant defence mechanisms genes from KEGG. Basically, KEGG identifies >> half of all genes with known COG nomenclature and notably higher than other functional databases.Thus I see GOMAPs exclusion of KEGG as a limitation. However, to reiterate to go identify COG/KEGG -> KEGG database interrogation and analysis is a lot of custom code.

Databases GOMAP implements Arabidopsis annotation, which may be cool and provides a starting point. It's also running UniProt, UniProt-plants, interpro, GO databases and those + Arabidopsis annotated DB could be really cool. It definitely sounds promising.

Processor time Any method of gene identification takes AGES and I have no doubt that GOMAP will be REALLY slow, looking at the pipeline it's using - there are a lot of databases its using. This is GOMAPs justification for using Singularity, it going to take ages to run therefore containerise it in a format often used by modern HPCs. To be honest it's a bit lazy, especially as it was published in 2021 (should be uploaded e.g. to conda). The authors talk about Python and R ... hmmmm, it looks to be a MySQL backend, thats where its speed will come from. The container will be large BTW hopefully is <1GB, but it will be >200MB.

Summary Its worth learning Singularity just keep to the basic commands to run GOMAP. You will need access to a large machine and the output will be okay but they're better general databases. However, I don't know how useful the Arabidopsis database it - thats about the biology, i.e. what you are looking for and whether Arabidopsis is a good model of that investigation

Note, Springer Nature refused a link image request for the container overview, but the image I wanted to share is here.

$\endgroup$
3
  • 1
    $\begingroup$ Just to mention the essential missing part the Singularity container test described singularity exec lolcow_latest.sif cowsay moo. The output is moo. Cows say "moo". $\endgroup$
    – M__
    Jun 23, 2023 at 2:21
  • 1
    $\begingroup$ Thank you for pointing to GOMAP, seems a good fit for my plan. Basically, I'm doing a plant database (already have an online demo) as my Master's final project and I wanted to add these detectors besides aggregating other data. I never heard about Singularity before, but, at base, I'm a software developer and I worked with docker, vagrant, virtualbox and other similar systems so it shouldn't be an issue to get it rolling. $\endgroup$ Jun 23, 2023 at 7:03
  • 1
    $\begingroup$ Thats good. The project makes a lot of sense. Singularity is huge in biological science (Docker uptake was low). The only thing to add is entering the Singular container is very different to Docker's -it. From there both the code and databases are open for inspection. $\endgroup$
    – M__
    Jun 23, 2023 at 10:03

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.