Aim 2.3A: Assessment of new knowledge algorithms

[schristley]

• [ ] What types of knowledge and/or algorithms do we want to compute?
• [ ] specificity prediction algorithms
• [ ] gather comprehensive list of existing tools
• [ ] which algorithms will require coding
• [ ] prioritize which tools to run, what are dependencies (data, software) are required
• [ ] Where will the algorithms run? HPC? GPU? AKC Server?
• [ ] more

[schristley]

• [ ] gather comprehensive list of existing tools

Gathering a list of tools in scalable KG google

[schristley]

• [ ] What types of knowledge and/or algorithms do we want to compute?

Similarity algorithms

• receptor similarity
• epitope similarity
• antigen similarity
• repertoire similarity

Specificity/binding prediction algorithms

• receptor-epitope specificity prediction
• receptor-antigen specificity prediction
• MHC-peptide prediction through IEDB service

Graph/Network analysis

• Network types: receptor-epitope, receptor-antigen, receptor-MHC-epitope, receptor-receptor, epitope-epitope
• Network types: multi-valued with disease, age, race/ethnicity, life events like immune exposure
• Network types: experimental data only, n-neighbors with computed similarity
• Connected components
• Degree distributions
• Centrality metrics: degree, closeness, betweenness, etc
• Network motifs, over-representation