2021-08-10 Meeting notes

Access to CUVPN

@Sebastian Raubach

• Unable to access CUVPN due to Cisco AnyConnect configuration

polypoloid dataset

@Evan Rees

@Yaw Nti-Addae

Awaiting decision on polyploid dataset

lettuce full dataset

@Former user (Deleted)

Is lettuce full genome too large to work with?

~200M markers / ~6B datapoints (markers * accessions).

Full dataset is 10’s of GB, probably not feasible to manipulate

Loading hapmap format will impact performance - genotype data are highly condensed

Difference between hapmap and VCF is large - not a fair comparison

Try ‘plain’ VCF with only GT fields

Need to account for conversion time from VCF to hapmap

Gigwa had limitation for number of markers to query?

data formats

@Former user (Deleted)

timings bcftools for slicing / processing / conversion

data export

@Pierre Larmande

need to specify when exporting VCF if annotations are present - will impact performance

parity / equivalence

@Pierre Larmande

when comparing datasets, need to account for annotations / info content in VCF

lettuce scaling

@Pierre Larmande

see if import times are linear across chromosomes

Measuring resource usage

@Pierre Larmande

Do we want to compare memory footprint?

Compare efficiency across platforms

CPU usage / multithreading

measuring time

@Evan Rees

Are we timing things consistently? Measuring resource usage