Some other (free) isoform deconvolution tools apart from Cufflinks:

RSEM http://deweylab.biostat.wisc.edu/rsem/

RPKMforgenes (this one also gives counts) http://sandberg.cmb.ki.se/rnaseq/

MISO http://genes.mit.edu/burgelab/miso/

rQuant http://www.fml.mpg.de/raetsch/suppl/rquant

Alexa-seq http://www.alexaplatform.org/alexa_seq/

NEUMA http://www.ncbi.nlm.nih.gov/pubmed/21059678

NSMAP http://www.biomedcentral.com/1471-2105/12/162

Flux Capacitor http://flux.sammeth.net/

"Next Generation Sequencing in functional genomics" Google this and click on the first option (The article from Werner T, briefings in bioinformatics). This article is quite useful one and will help you answering a lot of questions related to the role of NGS in functional genomics and ways of combining ChIP Seq and RNA-Seq data. Although, someone (here in forum) might also give you a ready made solution to your problem, but I will strongly suggest you to read this, for two reasons: 1) To get detailed review on various methods on integrating ChIP seq and RNA seq data. 2) To build your own hypothesis.

You may also have a look at the website of Broad institute, http://www.broadinstitute.org/cancer/software/genepattern/modules/RNA-seq/.

good luck.

The answer for first part of your question is cufflinks. "Cufflinks assembles transcripts, estimates their abundances, and tests for differential expression and regulation in RNA-Seq samples."

Second part is more tricky. What we successfully done so far, we selected up-regulated genes and looked for over-represented Transcription Factor Binding Sites (TFBS) in those. You need to compare promoters of up-regulated genes to all promoters in given genome. You can try RegionMiner: Overrepresented transcription factor binding sites from Genomatix. Probably you can do similar thing for genes having alternative splicing, but I have never tried that...