@Josh, thanks for the list -- I've defintely seen this one before and have worked my way through it. The problem is, much of the software listed there is over 10 years old, and several newly made scripts, such as Prodigal, are missing too. I'm more looking for people who do this often to report on their strategies
Automated Microbial Gene Prediction From Assembled Genomes -- What Is The Latest, Most Accurate Software?
I would like to get a feel for what sort of strategies other bioinformaticians use for conducting genome-wide gene prediction for prokaroytes. There doesn't seem to be much research into improving automated gene prediction for single organisms, with the latest articles concerning communities and metagenomics applications.
I am especially curious because these three methods all returned results that differed by about 10% when tested on the same genome. As I am new to gene prediction, I am interested to hear about other helpful software and in-house analysis techniques. How do these approaches differ when a reference annotated genome exists for training?
Here is the community-sourced list of software packages:
And web/literature sources:
Bacterial genome annotation systems
Please continue to add suggestions to this list, and I will update regularly as well!
4 answers
GenePRIMP from JGI was designed to solve your problem.
GenePRIMP stands for "Gene PRediction IMprovement Pipeline". The GenePRIMP pipeline consists of a series of computational units that identify erroneous gene calls and missed genes, and then correct a subset of the identified anomalous features. The data input to GenePRIMP needs to be a file of gene calls in GenBank or EMBL format. As its output, GenePRIMP generates reports of identified anomalies, plus a corrected EMBL file.
This blog (Bacterial genome annotation systems) also provides useful information.
In addition to whatever software you use to predict genes in the bacterial genome, make sure you run your sequence (translated in six frames with X as stop codons) against protein domain database such as Pfam. It helps to identify regions missed by gene calling software but also shows indels/false stop codons present due to sequencing errors.
• 216 views
•
link
written
by
Pawel Szczesny
Here's a good list of all the prediction programs out there. The difference you see in the programs is due to differences in motif calling. I've really only used AUGUSTUS (which works well IMO) but I'm fairly certain it uses Eukaryote motifs so I don't know how well it would work for bacteria and archaea (probably not at all).
+1, I knew I wasn't going to be much help, but I hope some other people report their experiences with gene prediction so we both can learn a bit. So far, AUGUSTUS has been good for me, but I would like to learn about more options including what new resources people are using.
Log in to answer this question.
More posts like this
-
Augustus and SNAP training in Maker-p pipeline
written by ashaneev07 •
-
Best Software For 454/Pacbio Hybrid Assembly
written by mgalactus
-
How To Determine Pcr Duplicate/Redundant Reads In Ngs Data?
written by JacobS
-
What Are Some Good Interactive Circular Genome Browsers?
written by Burke
-
Which Sequencing Simulator Is Better For Rna-Seq Simulation?
written by Tianyang Li •
-
Adapters Vs Pcr Fragments
written by ric •
-
What Is Your Experience With Gmod Tools Or Alternatives?
written by Michael Dondrup
-
Online Resources For Mouse Research
written by Andrew Su
-
Ngs, What'S New ? "I Am Preparing A Course On Ngs" , An Update For 2011/2012
written by Pierre Lindenbaum
-
How To Decide Which Software To Use For Building Automated Ngs Analysis Pipeline
written by Mcmahanl
