Search found 5 matches

by liyangum
Thu May 05, 2022 4:53 am
Forum: Main Forum
Topic: respre memory issue
Replies: 2
Views: 441

Re: respre memory issue

Hi,

Thank you for the feedback. Can you send us the MSA? ResPRE may use large memory with long sequences, but it should not reach up to 550G.
You can also send me the MSA by email.

Thanks,
Yang Li
by liyangum
Wed Oct 27, 2021 11:32 pm
Forum: Main Forum
Topic: ResPre precision matrix
Replies: 1
Views: 519

Re: ResPre precision matrix

Hi, Thanks for the interest. In fact, the plot in Fig.1 was at the size of LX21 by Lx21. You can plot the output matrix of ROPE in https://github.com/leeyang/ResPRE/blob/master/precision.py (LINE43) for the raw precision matrix for visualization. If you need an L x L matrix, you may check Text S4 of...
by liyangum
Thu Jun 17, 2021 2:30 pm
Forum: Main Forum
Topic: I have some questions about ResPRE
Replies: 1
Views: 673

Re: I have some questions about ResPRE

We did see the existence of imbalanced positive and negative samples. But we do not think that it is a problem.
Of course, you can set larger weights or other regular settings to try to make it more balance during the training.
by liyangum
Wed May 26, 2021 7:00 am
Forum: Main Forum
Topic: I have some questions about ResPRE
Replies: 5
Views: 1424

Re: I have some questions about ResPRE

We used supervised machine learning to train ResPRE models for contact map predictioon. The supervised information should be the contact maps (binary matrix, 1 for being in contact) from pdb structures.
by liyangum
Sun May 23, 2021 8:25 pm
Forum: Main Forum
Topic: I have some questions about the formula in the ResPRE paper
Replies: 2
Views: 1024

Re: I have some questions about the formula in the ResPRE paper

The calculation of frequencies with pseudo count has been used in many papers for coevolution analysis. For example, Eq 21 in plmDCA (https://arxiv.org/pdf/1211.1281.pdf), and Eq 1 in DCA paper (https://www.pnas.org/content/108/49/e1293).