Adobe Systems
Zápatí prezentace
1
Cufflinks
Peter Guman

Adobe Systems
Cufflinks
2
Cufflinks
̶Open-source software developed by laboratories at UC Berkeley and Johns Hopkins University
̶Cufflinks may be:
̶name of the RNA-Seq workflow: together with read mapper TopHat, tools Cuffmerge, Cuffdiff or
Cuffnorm
̶name of the standalone tool used for the assembly of mapped reads to assemble transcriptomes and
evaluate the expression
̶developed mainly in C++ and C languages allowing to have low-level control and optimized
performance

Adobe Systems
Cufflinks
3
What does Cufflinks do?
̶After the mapping phase of RNA-Seq workflow: assembly of different transcriptions/alternative
splicings
̶Takes input: aligned reads in SAM/BAM file format
̶Produces output:
1.Transcriptome assembly in GTF/GFF format. It has isoforms produced by Cufflink with 7 standard
GTF columns with optional information about genes and transcripts.
2.Transcript-level expression in FPKM Tracking file format contains isoform-level expressions.
3.Gene-level expression in FPKM Tracking file format on the gene level.
̶Optional: give the Cufflinks tool additional reference annotation. This way, Cufflinks can improve
already known references.

Adobe Systems
Cufflinks
4
Cufflinks algorithm
̶Creation of tree: Compatible fragments are connected when they are compatible*, and they have
overlap in the genome
̶'incompatible’ fragments must have come from differently spliced isoforms
̶isoform transcripts are generated, and an estimation of abundance follows. Estimation is based on
the assumption that the probability of observing each function is linear to the abundance of
transcripts of their origin.
̶
̶
*according the Dilworth's Theorem

Adobe Systems
Cufflinks
5
̶

Adobe Systems
Cufflinks
6
Assembler
̶Basic Cufflinks assembler creates Directed Acyclic Graph from the reads and finds the shortest
path between them (in a maximum parsimony fashion). Does not need a prior reference annotation.
̶reference annotation-based transcript (RABT) assembly is an improved algorithm. It requires
additional annotation of the already known genome.

Bioinformatics, Volume 27, Issue 17, 1 September 2011, Pages 2325–2329,
https://doi.org/10.1093/bioinformatics/btr355
The content of this slide may be subject to copyright: please see the slide notes for details.
Fig. 2. An overview of our RABT assembly method. First paired-end reads (mates shown connected by
solid lines) are ...

Fig. 2. An overview of our RABT assembly method. First paired-end reads (mates shown connected by
solid lines) are mapped to the genome using a spliced read mapper that can map reads across
junctions (shown in dotted lines). The reference annotation (blue) is used to generate faux-read
alignments that tile the transcripts (green). The faux-read alignments are used together with the
spliced read alignments to generate a reference genome-based assembly (black). This assembly is
merged with the reference annotation, and ‘noisy’ read mappings are filtered resulting in all
reference annotation transcripts in the output (blue) as well as novel transcripts (light blue).
Unless provided in the caption above, the following copyright applies to the content of this slide:
© The Author 2011. Published by Oxford University Press. All rights reserved. For Permissions,
please email: journals.permissions@oup.com© The Author 2011. Published by Oxford University Press.
All rights reserved. For Permissions, please email: journals.permissions@oup.com

Adobe Systems
Zápatí prezentace
8
Recources
̶Adam Roberts, Harold Pimentel, Cole Trapnell, Lior Pachter, Identification of novel transcripts in
annotated genomes using RNA-Seq, Bioinformatics, Volume 27, Issue 17, 1 September 2011, Pages
2325–2329, https://doi.org/10.1093/bioinformatics/btr355
̶Trapnell C, Williams BA, Pertea G, Mortazavi A, Kwan G, van Baren MJ, Salzberg SL, Wold BJ,
Pachter L. Transcript assembly and quantification by RNA-Seq reveals unannotated transcripts and
isoform switching during cell differentiation. Nat Biotechnol. 2010 May;28(5):511-5. doi:
10.1038/nbt.1621. Epub 2010 May 2. PMID: 20436464; PMCID: PMC3146043.
̶official site: cole-trapnell-lab.github.io/cufflinks/
̶