Address of the bookmark: https://www.sbgenomics.com/graph/

Careful encoding of graph entities allows odgi to efficiently compute and transform pangenomes with minimal overheads. odgi implements a dynamic data structure that leveraged multi-core CPUs and can be updated on the fly.

The edges and path steps are recorded as deltas between the current node id and the target node id, where the node id corresponds to the rank in the global array of nodes. Graphs built from biological data sets tend to have local partial order and, when sorted, the deltas be small. This allows them to be compressed with a variable length integer representation, resulting in a small in-memory footprint at the cost of packing and unpacking.

The RAM and computational savings are substantial. In partially ordered regions of the graph, most deltas will require only a single byte.

Address of the bookmark: https://github.com/pangenome/odgi

Though the specific implementation of the idea of the rectangle graph approach is already included into the current SPAdes distribution, we're also releasing the Rectangle Graph Module (RGM) as the separate code which can be run independently of SPAdes. Although RGM differs from the current implementation of the rectangle graph approach in SPAdes, in the future we plan to integrate RGM in SPAdes. RGM can be run with other genome assemblers if they use the graph format as SPAdes files.

For more details see: Nikolay Vyahhi, Son K. Pham, Pavel Pevzner. From de Bruijn Graphs to Rectangle Graphs for Genome Assembly, Lecture Notes in Bioinformatics 7534 (2012), pp. 249-261.

Address of the bookmark: http://bioinf.spbau.ru/en/rectangles

More at https://github.com/ekg/glia

glia -f ~/human_g1k_v37.fasta -t 20:62900077-62902077 -v variants.vcf.gz \
     -s AAATGTAAACATTTTATAGGGGATTCCCCTAAAAACAAAAAAACTTTCTGGGAAAGATTTTTCAAAAAATAAAA

Address of the bookmark: https://github.com/ekg/glia

• nodes, which are labeled by sequences and ids
• edges, which connect two nodes via either of their respective ends
• paths, describe genomes, sequence alignments, and annotations (such as gene models and transcripts) as walks through nodes connected by edges

Address of the bookmark: https://github.com/vgteam/vg

https://fungidb.org/fungidb/app/downloads/Current_Release/GultimumBR650/

Address of the bookmark: https://pangenome.github.io/

Requirements

You are (or soon will be) a master in bio-informatics. You have strong ICT skills and you are eager to fully submerge into the world of protein modeling. You have good experience with Linux and one or more programming languages as well as knowledge of tertiary structure analysis. Candidates with a Master degree in one of the life sciences (Biomedical sciences, Biochemistry, Bio-engineering, Biostatistics, …), with relevant interest and extended experience in this field are also welcome. A general background cancer biology and genetics is needed. You are willing and eligible to apply for a personal PhD fellowship with the Flemish FWO (FWO.be). Therefore, it is required that you hold a master degree from a European university, and have not obtained your master diploma more than three years ago (see FWO website for detailed conditions). Proficiency in English, and good communication skills, both oral and written, are required. You are highly motivated, and you like to work in an interactive research team. You are willing to work on a 4-year PhD project starting beginning of 2020.

What we offer

We offer a one year position, as a PhD student, which can be extended up to 4 year upon positive evaluation, even if a personal fellowship application is not successful. Wages are according to the standard Flemish bursary levels for PhD students.

Interested?
For additional information please contact dr. Geert Vandeweyer. To apply, send a copy of your CV including details of your relevant skills and a motivation letter by e-mail to dr. Geert Vandeweyer (geert.vandeweyer@uantwerpen.be) before March 15, 2020.

Source:https://academicpositions.be/ad/university-of-antwerp/2020/phd-student-bio-informatician-in-computational-protein-modeling/141252?utm_source=jooble&utm_medium=cpc&utm_campaign=jooble

The authors tested the pipeline across 5 datasets from 4 species, spanning an almost 80-fold difference in sample size (from 136 to over 10,700 individuals) and found that the predictive accuracy was within 0.02-0.05 of the standard GBLUP and Bayesian ridge regression approaches, again without the user having to manually select a model. More importantly the pipeline was able to recover known biological signals (FLC, FT and DOG1 in Arabidopsis, Waxy and GW5 in rice and Glu-D1 in wheat) and produced well calibrated association statistics. The most interesting observation though is that prediction accuracy and biological association are not necessarily the same thing. Two traits with identical predictive performance could have association significance that varied by over 60 orders of magnitude. So knowing how well we can predict a trait doesn't necessarily tell us anything about the distribution of the genetic architecture of that trait across the genome. That's where GenoBridge is actually of value. It's not just another machine learning pipeline. It's an example of an adaptive analytical workflow that doesn't just throw algorithms at the data but asks whether the data actually justify the next analytical step. As the amount of genomic data we generate continues to grow and the number of computational methods available to us increases, the bottleneck may not be in finding the right algorithm at all. The bottleneck may be in deciding which algorithm we can actually trust.

GenoBridge offers a potential solution: let the data tell you which model is appropriate, validate the prediction rigorously and only then move from prediction to biological discovery. That may be where computational genomics is heading. Rather than a future in which there is one single 'best' model, we may be heading towards a future of intelligent systems that know when a model is appropriate, when a result is reliable and when the biology is actually telling us something that we should be listening to.

More at https://link.springer.com/article/10.1186/s13007-026-01596-5

More at https://www.youtube.com/watch?v=vdGuy1GAJrQ

Address of the bookmark: https://sourceforge.net/projects/accnet/