deletions | additions
diff --git a/Genome Assembly and Annotation.md b/Genome Assembly and Annotation.md
index d72afe1..ca13e7d 100644
--- a/Genome Assembly and Annotation.md
+++ b/Genome Assembly and Annotation.md
...
There are three components to genome assembly validation. The first is the overall "quality" of the assembly, assessed by examining the stats provided by A5-miseq (discussed below). The second is verification that the organism sequenced is the organism of interest, simply by checking the assembled 16S sequence with BLAST. The third is "completeness" which is difficult to measure except in cases where a close reference is available. Nevertheless, we can get an idea of how complete the genome is by looking for highly conserved "housekeeping" genes that are found in almost every bacterial genome. To do this, we use a program called PhyloSift \cite{Darling_2014} to assess the presence or absence of 37 housekeeping genes in the assembly to infer completeness.
###Interpretation of A5-miseq stats
To open A5-miseq stats, import it into excel as a tab
deliminated delimited CSV file.
The first two numbers, shown in columns 2 and 3, are the number of contigs and scaffolds. Defining a "good" or "bad" assembly starts here. A finished assembly would consist of a single contig with no unresolved nucleotides but that is extremely unlikely to result from short read data. At the other extreme, we would consider a bacterial assembly in 1000 contigs to be very fragmented. In our experience, acceptable bacterial assemblies using
Ilumina Illumina PE300 data, assembled with A5, tend to range from 10-200 contigs. It is also worth noting that unless studying genomic organization, the number of contigs is less important than the gene content recovered by the assembly which is typically >99% using A5-miseq (Coil et al, submitted).
"Genome Size" and "Longest Scaffold" are simply represented as base pairs. While genome size can vary within taxa, this can be a second useful sanity check for the assembly. When expecting a 5MB genome based on other sequenced isolates from the same genus, if the assembled genome size is 2MB or 10MB, a red flag should be raised. "N50" represents the contig size at which at least 50% of the assembly is contained in contigs of that size or larger. This metric, combined with the number of contigs is the most common measure of assembly quality… larger is better. An N50 of 5,000 bp would be quite poor... meaning that half of the entire assembly is in contigs smaller than 5,000 bp. On the other hand an N50 of 1,000,000 bp is considered very good for bacterial genomes sequenced with Illumina technology.
...
From the PhyloSift directory
Move to the "PS\_temp" directory
Within this
directoy, directory, Phylosift has created a directory with the same name as the input file. Move to this new directory,
and then move to "blastDir".
Open the marker\_summary.txt file in the blastDir
...
There are a number of different pipelines available for annotation of bacterial genomes. These include Prokka \cite{Seemann_2014}, IMG \cite{Markowitz_2014}, RAST \cite{Overbeek_2014}, GLIMMER \cite{Delcher_2007}, PGAP \cite{Angiuoli_2008} and others.
Each of these pipelines has advantages and disadvantages, and each will give slightly different results. Here we recommend RAST since it is web-based, easy to use, returns results within hours, and provides a convenient toolbox for analyzing the results. However, RAST annotations are very difficult to submit to NCBI so we recommend allowing NCBI to re-annotate the genome with PGAP upon submission. Also, we recommend reporting the annotation results from the PGAP annotations in the genome announcement (for consistency.) Why do we also run a RAST annotation? Because we are impatient and we like to see results right away. We do not like having to wait for the NCBI
sumbission submission process to be completed before we start exploring our data.
###RAST Annotation
Navigate to http://rast.nmpdr.org/ and register a new account. Once you have created an account, log in.
Hover over the "Your Jobs" tab at the top of the page and click on "Upload New Job." In order to proceed you must specify a domain, a genus, a species, and the genetic code (usually "11".) Click "Finish the Upload."
The annotation will take some time, ranging from 2 hours to a few days, depending on server load. RAST will email you when it is complete. Once the annotation is complete, use their SEED Viewer to explore the annotation and metabolic pathways of the organism. From the RAST results, you can obtain information such as the
precense presence or absence of a particular gene/pathway and you can compare the annotation to other genomes in their database.