The input may be transcripts with optional read counts, proteins with optional spectral counts, or sequences to be assembled.

The Add panel allows you to specify the sequence dataset (fasta file) along with its conditions (e.g. tissue, treatment, etc), which will be listed under Associated Counts on the main panel.

If there are replicates, you may define them with the Define Replicates panel.

→ 2. Instantiate

Skip Assembly - the sequences can be instantiated with no assembly.

Assembly - the sequences, such as Sanger ESTs, 454 reads, and/or transcript, can be assembled.

Add/Edit: Selecting Add shows this panel. Selecting Edit shows a similar panel, but only the ATTRIBUTES values can be changed after the database has been created. The counts can be in one file or generated with the Build combined count file option. On Save, the condition names will be written in the Associated Counts table on the main panel.

Build combined count file file: (panel not shown) If the counts are in multiple files, this button opens up a panel in which a directory of count files, or individual count files, can be entered. It will generate a file called Combined_read_count.tsv where the columns are conditions with their respective counts.

Define Replicates: If the Associated Counts table has replicates, they can be defined in this panel, which will update the table as shown above in the Main panel.

Edit Attributes: (panel not shown) This panel has the same attributes fields as the Add or Edit a Sequence Dataset, but will be applied to the selected condition. These can be edited after the database has been created.

Add/Edit: Databases to search against are referred to as "annoDBs". Any FASTA file can be used an annoDB, though TCW gives special support to using UniProt taxonomic databases; besides allowing taxonomic specific querying in viewSingleTCW, the UniProt .dat file is used to extract GO, KEGG, EC and PFam information. The runAS program provides this functionality.

As the UniProt TrEMBL databases increase in size, it has become impractical to search these with BLAST¹. Fortunately, the diamond² programs provide super fast results.

ANNOTATIONS: An option of special interest is the Prune option. The results of the search against annoDBs can result in many hits, where some have the exact same alignment values and many have the very similar descriptions. The Alignment option removes all but the best hit with the same alignment and description values, and the Description option removes all but the best hit with same description values.

ORF FINDER:The ORF Finder uses the seq-hit frame to determine the best open reading frame. If a sequence has no hits or hits to multiple frame, it decides the best frame using the ORF length and the Markov score. The Markov scores are trained using the best hits over all sequences.

SIMILAR PAIRS: This is useful to determine how similar the input sequences are.