Project Proposal

chall84 final project proposal

Multianalysis VCF viewer

Input = VCF file annotated by SnpEff, VCF file annotated by ANNOVAR
(Can I have the user upload a VCF and then annotate it on the server? It would be ideal for user experience but if not I can just do the analysis on my own computer and copy the analyzed files to the server)

Files are parsed and loaded into the database. Each file is essentially its own table in the schema. Also in the database is a "genes" table parsed from the GFF3 for the human genome. This table links in gene ontology terms for filtering. The "dbSNP" table has rows for each rs# and links in the clinical significance. To save space I'll just download a list of the pathogenic or likely pathogenic snps for the table and assign a value of "benign" to any variant that has an rs# that's not in the table.

Schema:

SnpEff variants 
columns = variant id#(primary key), location, ref, alt, SnpEff consequence significance prediction, genecode(foreign key)

ANNOVAR variants 
columns = variant id#(primary key), location, ref, alt, rs#(foreign key), ANNOVAR consequence significance prediction, genecode(foreign key)

genes
columns = genecode(primary key), location, ontology

dbSNP
columns = rs#(primary key), location, clinical significance

The output on the webpage will be a table with columns= variant location, ref, alt, rs# (or novel), gene, gene ontology, SnpEff consequence prediction, ANNOVAR consequence prediction, dbSNP clinical significance.

The user will be able to filter the results table by the values within the columns, for example, a user might choose to view only those variants with a GO of DNA repair and a pathogenic predicted consequence from SnpEff or ANNOVAR. Or only those variants with an rs# labelled pathogenic.This allows the user to compare predictions to each other and to known consequences.