Revert parameter tuning commits

ksahlin / strobealign

Aligns short reads using dynamic seed size with strobemers

MIT License

141 stars 17 forks source link

Revert parameter tuning commits #358

Closed marcelm closed 10 months ago

marcelm commented 10 months ago

I’d like to suggest that we revert the parameter tuning from PR #345 and #351 until we have come to a conclusion about the problem with the reduced mapping rate discussed in #354.

This would allow us to make a release with the other fixes that we have accumulated since v0.11.0.

marcelm commented 10 months ago

I have run this PR on the evaluation datasets and compared it to v0.11.0.

As expected:

Mapping rate is nearly unchanged
Accuracy is nearly unchanged (average differences: +0.0027 SE and -0.0022 PE)

ksahlin commented 10 months ago

Approved (assuming the -0.0022 is percentage points, i.e. it changes about 2 reads in 100,000.)

marcelm commented 10 months ago

Yes, this is in percentage points.