LitSelect is the selection method associated our PALMER approach for gene-GO biclustering. PALMER was developed at the Medical University of South Carolina (MUSC). For any inquiries, please contact any members of the project, listed below.

• Daniel Couch (software), couchd@musc.edu
• Zhenning Yu (software), yuzhenning.bio@gmail.com
• Dr. Dongjun Chung (algorithm), chungd@musc.edu

The selection method used by LitSelect is as follows.

• For each gene list provided, a "candidate" gene list is selected. These genes are those most similar to the input list. Namely, they are the genes with the highest average cosine similarity to the genes in the input list.
• From the input gene lists, candidate GO terms are also selected. These are the GO terms with the highest average p-value among the genes under consideration.

LitSelect generates a binary gene-GO matrix using the selected genes, GO terms, and a user-provided cutoff for the p-value. If the p-value between a gene and GO term is under this cutoff, then the entry in the matrix receives a value of 0. If it exceeds the cutoff, it receives a value of 1.

LitSelect uses the backend database GAIL as its data source. The database has an asociated webserver for interactively visualizing gene-gene associations derived from literature mining.

We also provide an R package 'palmer', associated with the novel biclustering algorithm. In addition to implementing the algorithm, the package also makes the literature mining data accessible through an API to facilitate usage of the algorithm.

• GAIL
• PALMER

Coming soon...