AutoPlasmid was developed to take the process of annotating a plasmid to a whole new level. Firstly, we wanted to create a centralized depot of many thousands of well-known features and their associated sequences. We chose what we thought were some of the more popular features, and relied on databases such as SGD, PlasMapper, and the University of Wisconsin Madison's E. Coli Database. We aggregated over 40,000 features and compiled them into a complex database with the following tables: Annotation, Feature, FeatureType, Oligo, Organism, Pathogen, Plasmid, Registry Part, Registry Type
But we knew this database would grow as large as it did. To address this concern, we implemented Cloud Computing, what we feel is one of the most impressive aspects of our software application. A plasmid such as puc18 normally takes 39 minutes on a slow computer to annotate and search through all our features. But in the cloud, with 30 parallel processes running at once, we can split up the algorithm and divvy up the work, resulting in a run of just 68 seconds, a dramatic decrease in time by 34-fold!
But we added many, many more features on top of this architectural backend. In particular, we gave users the ability to annotate their plasmid using imperfect matches selecting any threshold they want (ex. 95% match). While this significantly increases the time a particular annotation takes, it still results in rapidly fast annotations when the Cloud is used. In addition, users may select particular features they want to search for, such as Genes, Promoters, Terminators, and more. Lastly, users may translate DNA into amino acid in all 6 reading frames.
Once a plasmid is fully annotated, the user can do even more. They can look inside the plasmid, view details about each feature that was annotated, and manipulate the plasmid even more. They can add custom annotations, view where a particular segment of DNA is, and even isolate out features to later design with. Users can view oligo matches, and amino acid translation of particular DNA segments. One of the most useful features of the AutoPlasmid's plasmid view window is that a user can find exactly why a annotation was identified as an imperfect match. In puc18, it is well known that the origin has a 1 base pair mutation, for example. This can be easily located by viewing the details of the alignment.
Lastly, a user may export their annotated plasmid in any format they'd like: genbank, fasta, or SBOL. These standardized file formats are useful when using AutoGene in collaboration with other standard softwares, such as Ape. Our genbank format, in particular, preserves characteristics of annotations that may be crossed over to Ape, such as the color of each annotation.
***NOTE: scroll down below for examples images showing screenshots of all the aforementioned features.
One of the key issues we recognized with the Autogene was the organization of the database. The structure would have to be robust to help encompass the different types of feature type annotations (terminator, promoter, coding sequence, etc.), organism taxonomy, and custom annotations made by the user. Therefore, there was a need to:
- standardize the different feature type names
- standardize organism taxonomy
- keep the database flexible and capable of managing new types of features
To this end, we looked for current standards for feature type annotations and organism taxonomic identification, and found the sequence ontology project as a reliable source for feature annotations and NCBI’s taxonomy database as a source of organism taxon identification.
We obtained gene sequences from the University of Wisconsin Madison's E. Coli Database, the Saccharomyces Genome Database, Virulence Factors of Pathogenic Bacteria, and the Registry of Standard Biological Parts.