First off, we need some data to test IsoEM2 on. Doing some quick searching, a reanalysis of some RNAseq done on some human neuronal progenitor cells.
The data can be downloaded using sra-tools (which can be installed with bioconda: conda install sra-tools).
The bash script in ./data downloads the files from SRP070895
I pulled the accession list and the run info from this website (boy howdy is the SRA hard to navigate).
h/t to the EdwardsLab for posting some nice documentation for using fastq-dump
IsoEM2 can be installed from their GitHub repo
Made a dir ~/tools to install isoem2 into
cloned repo into ~/tools
git clone https://github.com/mandricigor/isoem2.gitran the install shell script (make sure that make and java are installed):
cd isoem2
./install
Note: For Java, you need both the jre and jdk installed:
sudo apt install default-jre default-jdk
and running isoem2 generates it's help message
kallisto does pseudoalignments and can provide bam files. I'm curious to see if this is actually usable for downstream analysis that require transcript alignments.
install kallisto
conda install kallisto
cat'd together the cdna and ncrna and made an index out of them: Homo_sapiens.GRCh38.allrna.kallisto.idx
Note: there are paired and single end reads from this dataset