Like other indel callers, Scalpel works by performing de novo assembly of regions of interest, so that misalignment to the reference genome cannot obscure the presence of an insertion or deletion. Scalpel's innovation is to repeatedly check its assembly before comparing to the reference genome, to account for simple sequence repeats that are a regular source of error in indel calling. When Scalpel assembles an exon, it collects reads that map to that exon (including partial matches), splits them into k-mers, and creates a de Bruijn graph to span the exon; however, if it detects repeats in the map, it iteratively increases the size of the k-mers by one base until the repeats are eliminated. This ensures that the final assembly of the exon is highly accurate while minimizing compute time.

The Cold Spring Harbor team's validation of Scalpel, published over the weekend in Nature Methods, compares Scalpel's performance on a live whole exome against HaplotypeCaller and SOAPindel. The donor is an individual with serious neurological disorders, which may be linked to a high incidence of indels. One thousand indels from this individual's exome, called by one or more of the informatics pipelines, were selected for focused resequencing. This resequencing revealed a 77% true positive rate for Scalpel calls, dramatically better than the rates for either of the competing tools; Scalpel performed especially well with indels longer than five base pairs, a traditional weak point for indel callers.

Finally, the authors demonstrate Scalpel's use on a large set of genetic data from nearly 600 families who donated samples to the Simons Simplex Collection, a project of the Simons Foundation Autism Research Initiative. Scalpel found a very high enrichment for indels in children affected by autism, compared with their unaffected siblings, a pattern that persisted even after excluding common variants.

Address of the bookmark: https://github.com/andrewgull/MGERT

This software implements a de novo, alignment free measure of sample genetic dissimilarity which operates upon raw sequencing reads. It is able to calculate the genetic dissimilarity between samples without any reference genome, and without assembling one.

Address of the bookmark: https://github.com/kdmurray91/kwip

More at https://pubmed.ncbi.nlm.nih.gov/34951622/

Address of the bookmark: https://github.com/EbmeyerSt/GEnView

Applications are invited from candidates for the temporary positions on contractual basis for academic arrangement in the School of Biotechnology, SKUAST-J:

3 Assistant Professor (Bioinformatics) (01)

Rs. 25000/- per month for Ph.D. and Rs. 20000/- per month for M. Sc. (Fixed)

High second class M.Sc. degree (55 % marks) in Biotechnology/ Bioinformatics

NET in concerned subject/ Ph.D. & Teaching/ research experience in Bioinformatics/ Computational Biology as evident from documents /publications

The application should reach the office of Coordinator, School of Biotechnology, SKUAST-J, Chatha, Jammu-180009 (J&K) along with attested copies of the certificates on or before 10.01.2019 through registered post only.

More Info: http://skuast.org/site/Templates%20HTML/jobs/2018/ap-sbt-21-12-2018.pdf

Perl script for creating a publication-ready vectorial genetic/linkage map in Scalable Vector Graphics (SVG) format. The resulting file can either be submitted for publication and edited with any vectorial drawing software like Inkscape and Abobe Illustrator(R).

The input file must be a text file with at least the marker name (ID), linkage group (LG) and the position (POS) separeted by tabulations. Additionally a logarithm of odds (LOD score) can be provided. Any extra parameter will be ignored.

map.tsv

ID<tab>LG<tab>POS<tab>LOD
13519  12     0       0.250840894
2718   12     1.0     0.250840893
11040  12     1.6     0.252843341
...

https://github.com/pseudogene/genetic-mapper

Address of the bookmark: https://github.com/pseudogene/genetic-mapper

Address of the bookmark: https://bioinformaticsworkbook.org/dataAnalysis/GenomeAssembly/GeneticMaps/creating-genetic-maps.html

DeepVariant is an analysis pipeline that uses a deep neural network to call genetic variants from next-generation DNA sequencing data. DeepVariant relies on Nucleus, a library of Python and C++ code for reading and writing data in common genomics file formats (like SAM and VCF) designed for painless integration with the TensorFlow machine learning framework.

https://ai.googleblog.com/2017/12/deepvariant-highly-accurate-genomes.html

https://www.biorxiv.org/content/10.1101/092890v6

Address of the bookmark: https://github.com/google/deepvariant

Address of the bookmark: https://portal.nersc.gov/genomad

Documentation can be found at https://nanoporetech.github.io/katuali/.

Address of the bookmark: https://github.com/nanoporetech/katuali