Implemetation of Kabsch algoritm for finding the best rotation matrix --------------------------------------------------------------------------- x

General function to number a sequence and divide it into different regions @param sequence: The sequence string @param state_vecto

Place gaps when a number is missing. All except wolfguy are continuously numbered

Args: x: (..., num_heads * num_channels) Returns: (..., num_heads, num_channels)

Given a length of a cdr3 give back a list of the annotations that should be applied to the sequence. This function should be depreciated

set coordinates of masks following a unified distribution between the two ends

The main function for anarci. Identify antibody and TCR domains, number them and annotate their germline and species. It is advised to use

* @brief Helper function to close a file descriptor. * * Inspects @a fd and calls <b>close</b>(3) if it has a non-negative value. * *

1, collect those residues with dis<d; 2, calculate TMscore

Set the regions to be used. Will clear anything added using add regions.

split a long string into vectors by whitespace * line - input string * line_vec - output vector * delimiter - delimiter */

:param data: [n_edge, *dimensions] :param segment_ids: [n_edge] :param num_segments: [bs * n_node]

strip white space at the begining or end of string */

Get a dictionary which tells you the index in sequence 2 that should align with the index in sequence 1 (key) ali1: ----bcde-f---

Symmetrically number a CDR loop (e.g. CDRL1/CDRH2 for IMGT) @param length: Define the length of target CDR @param maxlength: Defin

Return the structure. Variable domains are annotated with the chosen numbering scheme. o Receptor chains are numbered with the desired number

Description: Convert global (external) coordinates q to local (internal) coordinates p. p <- R^{T}(q - t) Args: R: (

Description: Convert local (internal) coordinates to global (external) coordinates q. q <- Rp + t Args: R: (N, L, 3,

Args: logits: Logit matrices, (N, L_i, L_j, num_heads). mask: Masks, (N, L). Returns: alpha: Attention weights.

Outputs to open file If sequence_id is specified as an integer then only this sequence will be printed. Otherwise all sequences will be

Args: center: (N, L, 3), usually the position of C_alpha. p1: (N, L, 3), usually the position of C. p2: (N, L, 3)

Write numbered sequences to csv files. A csv file is written for each chain type. Kappa and Lambda chains are written to the same file

read list of entries from 'name' to 'chain_list'. * dir_opt is the folder name (prefix). * suffix_opt is the file name extension (suffix_opt). *

Get the partially matched sequence identity between two aligned sequences. Partial in the sense that gaps can be in the state_sequence.