Algorithm

Chaos Game Frequency Matrix Representation

• We installed a library (kaos)
• Got 1000 references and 1000 AMR e. coli. genomes (.fa)
• For each of the .fa files, we do the following
• We read one .fa file at a time and replace anything other than A, T, G, and C with a random 20-length string of ATGC (alternatively, you can try replacing with one random character)
• Create FCGR for a specific K (you may consider normalizing by N-K+1 and see if that makes sense)
• Sum all FCGRs across genomes of one of the Ref or AMR categories
• Try to plot the vector form of the matrices as a scatter between the two categories

Chaos game plot probability for different kmer length against different class

• Install kaos libary (https://cran.r-project.org/web/packages/kaos/index.html)
• Read 150 fasta sample from each category ( Susceptible, AMR, and Reference).
• Find the FCGR matrix for each sample and normalise to 1.
• Find the average of all the sample for a given category ( Susceptible, AMR, and Reference).
• Plot probability of AMR vs ref, Susceptible vs ref and AMR vs Susceptible for different kmer length (Inter group).
• Plot probability of AMR vs AMR, Susceptible vs Susceptible and Ref vs Ref for different kmer length (Intra group)

Chaos game plot absolute difference of different kmer against different class with 10 samples

• Install kaos libary (https://cran.r-project.org/web/packages/kaos/index.html)
• Read 10 fasta sample from each category ( Susceptible, AMR, and Reference).
• Find the FCGR matrix for each sample(do not normalise to 1).
• Find the average of all the sample for a given category ( Susceptible, AMR, and Reference).
• Plot top 100 kmer who has highest absolute difference between AMR vs Susceptible frequency (for varying kmer length).

Materials

1. Chaos game and genome
2. First paper - https://academic.oup.com/nar/article-abstract/18/8/2163/2383530?redirectedFrom=fulltext
3. blog - https://towardsdatascience.com/chaos-game-representation-of-a-genetic-sequence-4681f1a67e14
4. Bioinfo paper (2001) - https://academic.oup.com/bioinformatics/article/17/5/429/277428?login=true
5. Kaos R package paper (2019) - https://academic.oup.com/bioinformatics/article/36/1/272/5521624
6. Nice review (2021) - https://www.sciencedirect.com/science/article/pii/S2001037021004736
7. Data structure chaos game - https://almob.biomedcentral.com/articles/10.1186/1748-7188-7-10
8. Alignment free genome comparison wiki page (includes info theory stuff) - https://en.wikipedia.org/wiki/Alignment-free_sequence_analysis#cite_note-52
9. Excellent application in AMR - https://academic.oup.com/bioinformatics/article/38/2/325/6382301
10. Global alignment - https://bmcbioinformatics.biomedcentral.com/articles/10.1186/1471-2105-7-243