Based on collected contigs and a Hi-C signal, it is necessary to assemble the genome, that is, to orient and order the contigs and estimate the distances between them.

• Randomly choose the initial state (orientation or order)
• Repeat iteratively
  • Making a random change (changing orientation or order)
  • If the likelihood function (P) has increased, then we accept the change, otherwise we accept with probability: P(new state)/P(old state)

• Run main_orientation.py to orient contigs.
• Run main_ordering.py to order contigs.
• Run gap_size.py to simulate gap size estimating.

• Orientation chr2 contigs (length > 25k) plots of accuracy and log_likelihood during MCMC running:

• Python 3.7

Python dependencies can be installed with pip:

pip install -r requirements.txt

Data has to contain information about Hi-C reads in to follow a format

1. pairs.txt

| * | name of contig which contains the first piece of read | position first piece of read in contigs | name of contig which contains the second piece of read | position second piece of read in contigs | * | * |

2. len.tsv

| name of contig | its length |

3. layout.txt has to contain order of contigs with the current orientation

(see data)

1. MCMC algorithms http://statweb.stanford.edu/~cgates/PERSI/papers/MCMCRev.pdf

Theses

Presentation