Back

Transcriptome computational workbench (TCW): analysis of single and comparative transcriptomes

Soderlund, C. A.

2019-09-26 bioinformatics
10.1101/733311 bioRxiv
Show abstract

De novo transcriptome sequencing and analysis provides a way for researchers of non-model organisms to explore the differences between various conditions and species. The results are typically not definitive but will lead to new hypotheses to study. Therefore, it is important that the results be reproducible, extensible, queryable, and easily available to all members of the team. Towards this end, the Transcriptome Computational Workbench (TCW) is a software package to perform basic computations for transcriptome analysis (singleTCW) and comparative analysis (multiTCW). It is a Java-based desktop application that uses MySQL for the TCW database. The input to singleTCW is sequence and optional count files; the computations are sequence similarity, gene ontology (GO), open reading frame (ORF), and differential expression (DE). TCW provides support for searching with the super-fast DIAMOND program against UniProt taxonomic databases, though the user can provide other databases to search against. The ORF finder uses hit information, 5th-order Markov models and ORF length. For DE and GO enrichment, TCW interfaces with the R environment and an R script, where R scripts are provided for popular methods. The input to multiTCW is multiple singleTCW databases; the computations are homologous pair assignment, pairwise analysis (e.g. Ka/Ks) from codon-based alignments, clustering (bidirectional best hit, Closure, Best Hit, OrthoMCL, user-supplied), and cluster analysis and annotation. Both singleTCW and multiTCW provide a graphical interface for extensive query and display of the data and results. Example results are presented from two rhizome and one non-rhizome plant, where one of the rhizome plants has replicate count data from four tissues. The supplement describes how to reproduce all tables and figures. The TCW V4 software is freely available at https://github.com/csoderlund/TCW; the package contains the jar files, external software, and demo files.

Matching journals

The top 5 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.