Hello sharontan,
thank you for interest in our project!

The answer depends on what task you are solving.

The MUSCLE and Kalign are the tools intended for Multiple Sequence Alignment.  Multiple sequence alignment finds the "optimal" alignment between all of the provided sequences. In case of multiple sequence alignment KAlign works faster with large sequences then Muscle or Clustal.

The Bowtie (and also Ugene genome aligner) solves another task: it aligns all provided sequences to single reference sequence (such as genome for example) one by one. This process is known as sequence assembly. A pairwise alignment is performed between reference and each short read. Unlike multiple sequence alignment in sequence assembly all short reads are aligned independently from each other.

The primary goal of Unipro UGENE project is the integration of most popular computational biology analysis methods within single user-friendly visual interface. UGENE is a free software. Full feature list can be found at the UGENE web-page. There is also a presentation available on the web, feel free to check it out.  

Concerning Gmap: you are quite right, it is applied for EST (cDNA) and mRNA sequence alignment, while Bowtie is a "general DNA" short read aligner. Also Gmap performs genome mapping. However, I suppose Bowtie can be used for your alignment task. It is a memory-efficient and very fast short-read aligner. You can read more about Bowtie here.

To use Bowtie from UGENE, activate Tools->DNA assembly dialog from main menu. Select Bowtie as assembly method, then set path to your reference genome and your cDNA shortreads. Additionaly you can use a prebuilt-index for genome (there is an optional check box). The pre-built indecies are available from official Bowtie ftp.
Don't forget to read UGENE manual pages devoted to Bowtie.

Btw UGENE user manual is available as a printable PDF file. It's the best source of information about UGENE.

In my opinion it should be the other way. Ask them to give you a very good argument, and then come back to try to convince the Ugene authors to implement a plugin with those functions. They have enough work developing the software to be also explaining things you can find in the manual or defending algorithms.
Moreover, in the T-Coffee Home page the say: "You can use T-Coffee to align sequences or to combine the output of your favorite alignment methods (Clustal, Mafft, Probcons, Muscle...)".
You can use both, Clustal2 and Muscle from Ugene, so I don't see a big advantage. And please, read the steps I have to follow to use that in windows. If I had to do that for every function I use from Ugene...
Concerning the Kalign, probably its easy to implement the newer version for the algorithm. BTW, its interesting that they use features to improve the alignment.