Crowdly

acidic proteins

The basic units of chromatin, the nucleosomes, contain which components?

Histones

Proteins

Autosomes

RNA

DNA

In human, DNA methylation can be established and maintained by

Dmm1

Dnmt1

Dnmt3a

Dnmt2

Dnmt3b

Splicing is often required for the maturation of this enzyme’s products.

Pol II

50%✨

Pol IIa

50%✨

Pol IV

50%

Pol V

Pol I

What are the “tokens” used to train large language models (LLMs)?

None of the other answers

Nodes within an artificial neural network

Word or sub-word pieces collected for use as training data

Digital currency

Free credits provided by Google or Amazon webservices

RNA polymerase IV and RNA polymerase V...

are completely redundant with Pol III

have no known biological functions

are completely redundant with Pol I

are completely redundant with Pol II

are recruited to silence transposable elements

Molecular biologists can use AlphaFold protein structural prediction to...

predict the interactions between two or more proteins in a complex.

None of the other answers

completely replace experimentation and close their wet laboratories.

investigate molecular machines with known function but unknown structure.

predict subunits of multisubunit complexes and match these to cryo-EM data.

RNA polymerase IV has a specialized function...

None of the other answers

in the biogenesis of messenger RNAs (mRNAs)

in the biogenesis of transfer RNAs (tRNAs)

in the biogenesis of small interfering RNAs (siRNAs)

in the biogenesis of microRNAs (miRNAs)

What are the “parameters” included in the architecture of typical LLMs?

None of the other answers

the number of artificial neurons

the number of weights and biases

the number of tokens used to train the network

the total number of neural network layers

What dataset(s) did Google use to train the AlphaFold2 Evoformer model?

> 250 million protein sequences and > 100 000 experimental PDB structures

Any structural data published on NCBI or World of Science

100 representative protein structures

Output from 500 000 protein molecular dynamics simulations

Publications from structural biology labs, interpreted by a large language model