Asteroid aborts: Assertion `newScore > _lastScore' failed

[LPDagallier]

Hi,

I'm facing an error when running Asteroid with multi-copy gene trees.

For example, running: asteroid -i test.trees -p asteroid_test -b 100 > asteroid.log I get the following error: "asteroid: /home/chrisribe/b/code/asteroid/src/core/DistanceMethods/AsteroidOptimizer.cpp:146: bool AsteroidOptimizer::computeAndApplyBestSPR(): Assertion `newScore > _lastScore' failed. Aborted (core dumped)"

In this case I don't use a mapping file because I renamed all the sequences names so that sequences from a same species have the same name. Asteroid is running on a cluster.

The test dataset and logfile are attached. test.trees.txt asteroid.log

[BenoitMorel]

Hi Léo-Paul,

Thanks a lot for the report. I also reproduce on my laptop. I think it's because you are not using a mapping file despite having multi-copy families (it's a bit hidden in the documentation, but see https://github.com/BenoitMorel/Asteroid#mapping-file). Basically, Asteroid can't read trees with duplicated names.

Sorry, I should have added a check to detect this and report an explicit error. Also, now that I think about it, I could also support what you are doing (duplicated names that represent the same species). I'll try to do it next week, so if you are not in a hurry, you can also wait for my update.

Have a nice week-end, Benoit

Le ven. 7 juil. 2023 à 20:19, Léo-Paul Dagallier @.***> a écrit :

Hi,

I'm facing an error when running Asteroid with multi-copy gene trees.

For example, running: asteroid -i test.trees -p asteroid_test -b 100 > asteroid.log I get the following error: "asteroid: /home/chrisribe/b/code/asteroid/src/core/DistanceMethods/AsteroidOptimizer.cpp:146: bool AsteroidOptimizer::computeAndApplyBestSPR(): Assertion `newScore > _lastScore' failed. Aborted (core dumped)"

Asteroid is there running on a cluster.

The test dataset and logfile are attached. test.trees.txt https://github.com/BenoitMorel/Asteroid/files/11984789/test.trees.txt asteroid.log https://github.com/BenoitMorel/Asteroid/files/11984784/asteroid.log

— Reply to this email directly, view it on GitHub https://github.com/BenoitMorel/Asteroid/issues/7, or unsubscribe https://github.com/notifications/unsubscribe-auth/ADJJ3UP66WG35ZT3LYDFHDDXPBHL7ANCNFSM6AAAAAA2CEIYAY . You are receiving this because you are subscribed to this thread.Message ID: @.***>

[BenoitMorel]

Hi Léo-Paul, If you update Asteroid (from the repository: git pull && ./install.sh), your run should now work. Let me know if not :) Best, Benoit

Le sam. 8 juil. 2023 à 00:00, Benoit Morel @.***> a écrit :

Hi Léo-Paul,

Thanks a lot for the report. I also reproduce on my laptop. I think it's because you are not using a mapping file despite having multi-copy families (it's a bit hidden in the documentation, but see https://github.com/BenoitMorel/Asteroid#mapping-file). Basically, Asteroid can't read trees with duplicated names.

Sorry, I should have added a check to detect this and report an explicit error. Also, now that I think about it, I could also support what you are doing (duplicated names that represent the same species). I'll try to do it next week, so if you are not in a hurry, you can also wait for my update.

Have a nice week-end, Benoit

Le ven. 7 juil. 2023 à 20:19, Léo-Paul Dagallier @.***> a écrit :

Hi,

I'm facing an error when running Asteroid with multi-copy gene trees.

For example, running: asteroid -i test.trees -p asteroid_test -b 100 > asteroid.log I get the following error: "asteroid: /home/chrisribe/b/code/asteroid/src/core/DistanceMethods/AsteroidOptimizer.cpp:146: bool AsteroidOptimizer::computeAndApplyBestSPR(): Assertion `newScore > _lastScore' failed. Aborted (core dumped)"

Asteroid is there running on a cluster.

The test dataset and logfile are attached. test.trees.txt https://github.com/BenoitMorel/Asteroid/files/11984789/test.trees.txt asteroid.log https://github.com/BenoitMorel/Asteroid/files/11984784/asteroid.log

— Reply to this email directly, view it on GitHub https://github.com/BenoitMorel/Asteroid/issues/7, or unsubscribe https://github.com/notifications/unsubscribe-auth/ADJJ3UP66WG35ZT3LYDFHDDXPBHL7ANCNFSM6AAAAAA2CEIYAY . You are receiving this because you are subscribed to this thread.Message ID: @.***>

[LPDagallier]

Hi Benoit, Oh that is great ! Thanks a lot for implementing this feature so fast. I'll try on my side and let you know. Best, Léo-Paul

[LPDagallier]

It works perfectly, thanks a lot Benoit !