ibnu2026

Notes

• Getting a lightning-specific error with hisat2

Aligner Test

Each aligner runs in its own environment, which is in the envs directory. Install them all as follows:

cd envs
ls | xargs -I {} conda env create -f {}

At the time of testing, all of the aligners have the same version number in both macOS-arm64 and linux-64 except bbmap.

Program	macOS-arm64	linux-64	Notes
bbmap	39.77	39.79
blast-legacy	2.2.26	2.2.26
bowtie2	2.5.5	2.5.5
bwa	0.7.19	0.7.19
gmap	2025.07.31	2025.07.31	broken
hisat2	2.2.2	2.2.2
minimap2	2.30	2.30
pblat	2.5.1	2.5.1	semi-broken on MacOS
segemehl	0.3.4	0.3.4
star	2.7.11b	2.7.11b	broken on MacOS
subread	2.1.1	2.1.1

Accuracy Experiments

Sequence sources

Name	File	Notes
random	random.fa.gz	created with randomseq.py
genome	at1pct.fa.gz	from init/data
mRNA	at1pct-mRNA.fa.gz	from extract-transcripts.py

Read sources

• Randomly generated reads decorated with 0-20% error
• Mock reads created by badread (not done yet)

Data for short/long reads and 3' ends

1. Use the setup.yml to create the setup environment.
2. Make sure grimoire is in PYTHONPATH
3. Make sure grimoire/bin/haman is in PATH
4. Run dataprep.py on a FASTA and GFF

The result is a file of semi-unique genes, each with a single, canonical/longest mRNA. Run dataprep.py like this:

python3 dataprep.py at1pct.fa.gz at1pct.gff3.gz AT1

These all end up in a files called

• AT1-genome.fa mini chromosome
• AT1-genome.gff3 gff for mini chromosome
• AT1-mRNA.fa transcripts only