This comment will serve as my notes for the article:

Methods
-Raw RNA-seq reads were pre-processed using trimmomatic with default parameters
-RNA-Seq read quality was assessed using FastQC and aggregated using MultiQC
-Trinity was used for de novo transcriptome assembly and validation was done using BUSCO
-They used HISAT2 to map reads to respective references
-Used TransDecoder with default setting to predict coding regions, they selected the best open reading frame per transcript longer than 100 peptides
-Used CD-HIT-EST with default params to reduce redundancy and produce unique genes
-Used BLAST to assign function annotations to the unique genes

-BLASTx program was used to annotate genes based on UniProtKB, which is a manually annotated, non-redundant protein sequence db.

How can we use this on Tomatoes
-Could go back and try validating our assemblies with BUSCO
-Try running TransDecoder on our contigs to predict coding regions, perhaps after finding the best long hits from BLAT
-Use CD-HIT to find our unique genes
-Use BLASTx to annotate our genes

I think we can dismiss BUSCO in this instance.
Since we are only looking at a transcriptome in some pistil tissue and maybe some pollen.
For BUSCO and completeness, we'd want leaf, and root and flower and all of the tissue sequenced to assess completeness.

Transdecoder, let's do it!!

CD-HIT, Heck yeah!!
These will be new tickets.