search

MissionBio / mosaic

Tertiary analysis package to visualize single-cell data generated by the Tapestri Platform.
17 stars 6 forks source link

Differences in variants between v1.8.0 and v2.0.4 #10

Closed jil24 closed 2 years ago

jil24 commented 2 years ago

I have a DNA+Protein dataset composed of 7 merged samples that I had previously been analyzing with mosaic version 1.8.0.

As I am trying out version 2.0.4 I find that the vast majority of the variants that I know are in the dataset do not seem to be recognized; before any filtering is applied.

The difference is quite dramatic:

Version 1.8.0

print(len(Merged_Samples.dna.ids()))

Result: 65151

Version 2.0.4

for sample in Merged_Samples.samples:
    print(len(sample.dna.ids()))

Result: 44 44 44 44 44 44 44

Is there some kind of implicit filtering going on? If so, how can it be disabled (I know that some variants have been filtered out that are real and confirmed with orthogonal assays)

KKJSP commented 2 years ago

@jil24 The default value of the apply_filter parameter in missionbio.mosaic.io.load has been changed from False in v1.x to True in v2.x. All the variants will be loaded when apply_filter=False is passed.

jil24 commented 2 years ago

Perfect, thank you