PanRNA
A useful tool for the complete transcriptome analysis of non-reference microorganism strains. It is based on a hybrid approach between reference mapping and de novo assembly of unmapped/unnanotated reads.
​
Check it out on Github!
Ancestral reconstruction state of large phylogenies
Tools like treetime and iq-tree have large memory requirements when reconstructing phylogenies ancestral states for large alignments. We developed a method to overcome that issue by dividing an alignment previously used to make a phylogenomic tree into chunks.
Predicting drug resistance in M. tuberculosis by Image Recognition NN
I used a transfer learning approach from the well known resnet-18 designed for image recognition, to adapt this neural network to classify isolates of M. tuberculosis into resistant of sensitive to Rifampicin based on genomics information.
You can try your own sample with the google colab code!