A direct database connection would allow to save annotations, blast-results, etc. directly into a local database

Blast2GO would save the actual project e.g. every 30 minutes

Improved pie charts with the possibility to dynamically select the information shown in pie charts

Dynamic data-export function: a graphical interface to select which type of information to be exported and in which order

Some organisms use non-standard codon table. So if we have an option
to select the codon table numbers for different organisms, it could be
better. I think it could be easy to incorporate this step in the tool
for translated query searches like blastx.

More interactive graphs and personalized graphs to adjust colours and shapes depending on functional groups or statistical values

A thing that would be useful for B2G is the ability to export the Blast results without mapping eg the same as 'Export/ Export Mapping Results/ *.txt-TSV-by Seqs' except just for Blast results

When you ask Blast2GO to draw a graph, it also lists the raw output
numbers on the application messages screen. So, you can copy those numbers to a separate graphing software of your
like to recreate the image.

Programmable Application Interface (API) to Blast2GO which would allow to access all the different options in Blast2GO in a programmable way

More advanced KEGG pathway functions (colouring, statistical tests, data-export)

As far as I can tell, unlike the GUI version, when using the stand-alone pipeline program, you give it a blast (and optionally IPR results) file but not a sequence file. The dat file that is created does not contain any sequence. Thus it becomes difficult to do some of the GUI functions. Being able to save the sequence in the dat (akin to the GUI method) would be great.

Because the MIPS categories of Arabidopsis genes are very useful for other plant genes

More advanced InterProScan functions: start/stop remote queries, access to a local server, save results as html, etc.

It would be very useful to have a search-and-replace function to remove/change GO terms or descriptions in multiple records at once. For instance I had to remove "auxin biosynthetic process" in 300+ records of a fungal (do not produce auxins) annotation project

I ask the B2G could provide detail alignment results like which ORFs are aligned with my sequence reads.

HUGO gene identifier for the best human homolog of your sequence. Not sure how practical this would be for non-vertebrates, but would feed nicley into eg Ingenuity Pathway Analysis

Mark sequences (with color?) that are unmappable, so that they can be removed from future mapping attempts. Otherwise there is no way to know which of the sequences were attempted to be mapped but couldn't.

Create a report package that allows you to automatically save selected graphs and tables. Will allow you to quickly visualize and save results outputs