I am new to bioinformatics and have been assigned the task of discovering which isoforms of a certain protein-coding gene are present in the RNA-seq data that I have. There are several isoforms of this gene listed on Entrez & Ensembl and I want to know which of these isoforms are expressed in the RNA-seq data that I have been given. I am not sure how to begin to tackle this issue so If anyone has any ideas or advice, it would be greatly appreciated.
Tools like kallisto or salmon can perform a process call pseudo-alignment that matches reads from RNA-seq data to transcript isoforms present in a reference genome. You can then look at the output of these tools to see which isoforms for a given gene appear to be present in your RNA-seq data. These tools answer the question "how abundant is each transcript in my data?".
But if you want to answer a question like "are there differences in which transcripts are present between experimental conditions?", I'd look for tutorials on differential expression analysis, differential splicing analysis, or differential transcript usage.
If you do as @james-hawley rightly suggests, keep in mind that pseudo-alignment could detect a lot of transcripts with low TPM that are actually false positives, as reported here (it's a paper focused on scRNA-seq, but many statements are also valid for bulk). So, you may want to try to remove false positive transcripts, that will likely result with very low read count and TPM. Salmon's decoy function could help here.
As an alternative to pseudo-alignment, you could run StringTie on your BAM file to detect transcript isoforms, using a reasonable threshold of coverage over splice junctions.