by Dr. Pandurang Kolekar, Bioinformatics Engineer, Strand Life Sciences

Abstract:

Unique Molecular Identifiers (UMIs) are short random nucleotide sequences that are increasingly being used in high-throughput sequencing experiments. In this webinar, we will highlight the UMI-friendly features of Strand NGS v3.1 including support for handling well known and customised UMI libraries, QC metrics, consensus alignment, UMI-based family size filters for read list, genome browser enabled with UMI-specific features and filters, UMI-aware variant calling parameters, and exporting UMI-tagged aligned samples. These all features together empower users to harness the potential of UMI-tagged NGS data for deeper insights. A case study demonstrating application of these UMI-based features in Strand NGS for low frequency variant calling in cfDNA sample will be presented.

UMI-tagged NGS libraries allow, ultra-sensitive detection of low frequency variants from liquid biopsy samples using DNA-Seq and accurate quantification of transcript-level expression using RNA-Seq. The recent release of Strand NGS v3.1, is equipped with the necessary features to efficiently analyse UMI-tagged NGS data helping researchers and labs involved in rare variant calling like in cfDNA based cancer diagnostics, and accurate transcript quantification with RNA-Seq.

Webinar Details:

Session 1: 13 Dec 2017, 2:30 PM IST
Session 2: 13 Dec 2017, 9:30 PM IST

Register here: http://www.strand-ngs.com/webinar_registration

Address of the bookmark: http://www.sthda.com/english/wiki/rna-seq-differential-expression-work-flow-using-deseq2

What Are Kallisto and Salmon?

At their core, both Kallisto and Salmon are tools for quantifying transcript abundance from RNA-Seq reads. They bypass traditional alignment-based methods, replacing them with pseudoalignment or quasi-mapping, which drastically speeds up the process.

• Kallisto was developed by Lior Pachter’s lab and introduced the concept of pseudoalignment using a de Bruijn graph.
• Salmon, developed by Rob Patro’s group, builds on this idea with quasi-mapping and offers additional features like advanced bias correction.

Head-to-Head Comparison

1. Algorithm

• Kallisto uses pseudoalignment, focusing on matching k-mers from reads to a transcriptome index.
• Salmon uses quasi-mapping, which adds more flexibility and can also work with aligned reads (BAM files).

2. Input and Flexibility

• Kallisto works with raw FASTQ reads and requires a custom transcriptome index.
• Salmon accepts FASTQ or pre-aligned BAM files, giving you more workflow options.

3. Bias Correction

One of Salmon’s major advantages is its sophisticated bias correction system. It corrects for:

• Sequence-specific bias
• Positional bias
• GC-content bias

Kallisto offers basic sequence bias correction but lacks the comprehensive models found in Salmon.

4. Speed and Resources

• Kallisto is blazing fast and slightly more memory-efficient.
• Salmon is still very fast, but the added features can come at a small computational cost.

5. Output and Downstream Analysis

• Both tools provide transcript-level quantifications and support bootstrapping for variance estimation.
• Salmon can also summarize counts at the gene level if provided with a mapping file (--geneMap).
• Kallisto integrates seamlessly with Sleuth for differential expression analysis.
• Salmon works well with tximport, DESeq2, edgeR, and other Bioconductor tools.

Choosing the Right Tool

Example Command Lines

Kallisto (paired-end):

kallisto quant -i transcriptome.idx -o output -b 100 sample_R1.fastq sample_R2.fastq

Salmon (paired-end, bias correction):

salmon quant -i salmon_index -l A -1 sample_R1.fastq -2 sample_R2.fastq \
  -p 8 --validateMappings --seqBias --gcBias -o output

Conclusion

Both Kallisto and Salmon are exceptional tools that have transformed RNA-Seq analysis. Your choice largely depends on your priorities—whether it's speed, accuracy, flexibility, or compatibility with downstream tools.

For many users, Salmon offers a more complete and flexible solution, especially when bias correction and gene-level outputs are essential. However, Kallisto remains a favorite for quick, accurate quantification, especially when paired with the Sleuth pipeline.

• Bioconductor – A plethora of tools for analysis and comprehension of high-throughput genomic data, including 1500+ software packages. [ paper-2004 | web ]
• Biopython – Freely available tools for biological computing in Python, with included cookbook, packaging and thorough documentation. Part of the Open Bioinformatics Foundation. Contains the very useful Entrez package for API access to the NCBI databases. [ paper-2009 | web ]
• Bioconda – A channel for the conda package manager specializing in bioinformatics software. Includes a repository with 3000+ ready-to-install (with conda install) bioinformatics packages. [ paper-2018 | web ]
• BioJulia – Bioinformatics and computational biology infastructure for the Julia programming language. [ web ]
• Rust-Bio – Rust implementations of algorithms and data structures useful for bioinformatics. [ paper-2016 ]
• SeqAn – The modern C++ library for sequence analysis.

Distribution package for the Prgressive Cactus multiple genome aligner. Dependencies are linked as submodules

https://github.com/glennhickey/progressiveCactus

Address of the bookmark: https://github.com/glennhickey/progressiveCactus

https://sourceforge.net/projects/genobuntu/

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

https://github.com/TF-Chan-Lab/OMTools

Address of the bookmark: https://github.com/TF-Chan-Lab/OMTools

GROMACS conda https://bioconda.github.io/recipes/gromacs/README.html 

Address of the bookmark: http://haddock.science.uu.nl/enmr/services/GROMACS/main.php

More at https://github.com/snakemake/snakemake-wrappers

Address of the bookmark: https://snakemake-wrappers.readthedocs.io/en/stable/

Recently useR! conference have been organized University of Castilla-La Mancha, Albacete, Spain.

http://www.edii.uclm.es/~useR-2013//