SNPsplit: Allele-specific splitting of alignments between genomes with known SNP genotypes

May combine two claims.

The rationale for N-masking in allele-specific alignments is that mapping bias towards the reference allele is eliminated and both alleles of the same read get placed in the same position in the genome equally well.

“The rationale for N-masking in allele-specific alignments is that the mapping bias towards the reference allele is eliminated and both alleles of the same read get placed in the same position in the genome equally well.”

• cites [3] J Degner, 2009, Effect of read-mapping biases on detecting allele-specific expression from RNA-sequencing data., Bioinformatics. sourcePartly supported

  The Results (section 3; page not provided) say SNP masking eliminated the overall reference-allele bias in simulated and real data. But masking did not eliminate strong biases at individual SNPs: the paper explains that either allele can match another genomic location and map differently. Thus the claim that both alleles are placed at the same position equally well is not supported as a general outcome. (LLM, conf=0.89)

  Experimental — accuracy not yet measured

Probably not a substantial claim: may not pass S2 (checkable). May combine two claims.

The quality of genotype calls is crucial for allele assignment, so genotype data needs careful collection and quality control and filtering to avoid biases and false positive hits.

“The quality of the genotype calls is crucial for allele assignment, so the genotype data needs to be collected carefully and quality control and filtering is required to avoid biases and false positive hits 9.”

• cites [9] S Castel, 2015, Tools and best practices for data processing in allelic expression analysis., Genome Biol. sourceSupported

Probably not a substantial claim: may not pass S3 (load-bearing).

Bowtie2, BWA, HISAT2, STAR or any tool wrapping one of these aligners are confirmed to work for alignment to N-masked genomes.

“Software confirmed to be working for this approach include (but are not limited to) Bowtie2 11, BWA 12, HISAT2 13, STAR 14 or any tool wrapping one of these aligners.”

• cites [11] B Langmead, 2012, Fast gapped-read alignment with Bowtie 2., Nat Methods. sourcePartly supported

  The paper’s Abstract (page not provided) identifies Bowtie 2 as a read aligner and reports its use with a human reference genome, supporting its identification in the list. It does not say that Bowtie 2 was tested or confirmed to work with N-masked genomes; searches for masked bases and ambiguous reference characters found no such passage. The other aligners have separate citations in the citing sentence and are not assessed here. (LLM, conf=0.84)

  Experimental — accuracy not yet measured
• cites [12] H Li, 2010, Fast and accurate long-read alignment with Burrows-Wheeler transform., Bioinformatics. sourcePartly supported

  The Implementation section identifies BWA-SW as a component of the BWA program, supporting the citation’s identification of BWA as an aligner. The paper discusses alignment to reference genomes but does not establish that BWA works for alignment to N-masked genomes; searches for N-masking and ambiguous bases found no such test or statement. (LLM, conf=0.84)

  Experimental — accuracy not yet measured
• cites [13] D Kim, 2015, HISAT: a fast spliced aligner with low memory requirements., Nat Methods. sourcePartly supported

  The Results section (‘Design principles of HISAT,’ page unavailable) identifies HISAT as a genome aligner. But the cited paper describes HISAT, not HISAT2, and the passages do not establish that either works with N-masked genomes or that tools wrapping them do. The citation therefore supports the general aligner identification, not the claimed compatibility. (LLM, conf=0.86)

  Experimental — accuracy not yet measured

May combine two claims.

Standard Hi-C cannot discriminate whether an interacting fragment originated from the same or the other allele, while allele-specific interaction maps can separate cis-allele from trans-allele interactions, improving analysis of chromatin dynamics and gene regulation.

“While standard Hi-C cannot discriminate whether an interacting fragment originated from the same or the other allele, allele-specific interaction maps can separate cis-allele from trans-allele interactions, thereby greatly improving the analysis of chromatin dynamics and gene regulation 20, 21.”

• cites [20] J Dixon, 2015, Chromatin architecture reorganization during stem cell differentiation., Nature. sourcePartly supported

  Split into 2 atoms — 1 confirmed, 1 partially supported, 0 disputed, 0 misattributed, 0 unverifiable. Expand the claim to see per-atom verdicts.

  Experimental — accuracy not yet measured
• cites [21] S Rao, 2014, A 3D map of the human genome at kilobase resolution reveals principles of chromatin looping., Cell. sourcePartly supported

  Split into 2 atoms — 1 confirmed, 1 partially supported, 0 disputed, 0 misattributed, 0 unverifiable. Expand the claim to see per-atom verdicts.

  Experimental — accuracy not yet measured

Probably not a substantial claim: may not pass S3 (load-bearing).

Allele-specific methylation at differentially methylated regions is maintained in a parent-of-origin specific way at the Gnas/Nespas locus.

“Allele-specific methylation at differentially methylated regions (DMRs) is maintained in a parent-of-origin specific way at the Gnas/Nespas locus.”

• cites [6] W Xie, 2012, Base-resolution analyses of sequence and parent-of-origin dependent DNA methylation in the mouse genome., Cell. sourcePartly supported

  The Introduction (page not available) says that imprinted DMRs carry parent-of-origin-dependent allele-specific methylation and that methylation at some DMRs is maintained throughout development. The Results describe parent-of-origin-dependent DMRs at other loci, but the available passages do not establish this specifically at Gnas/Nespas; a search for those locus names found only a Gnas reference-list entry. (LLM, conf=0.82)

  Experimental — accuracy not yet measured

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

ZFP57 binding is methylation dependent and can be found exclusively on the maternal allele at the SNRPN locus.

“The binding of ZFB57 is methylation dependent and can be found exclusively on the maternal allele (genetic background of mother in forward cross: Black 6; mother in reverse cross: Cast).”

• cites [17] R Strogantsev, 2015, Allele-specific binding of ZFP57 in the epigenetic regulation of imprinted and non-imprinted monoallelic expression., Genome Biol. sourceSupported

May combine two claims.

At the Gnas/Nespas locus, the upstream DMR is methylated exclusively on the paternal allele, while the more downstream DMR is methylated exclusively on the maternal allele in both forward and reverse crossed hybrid mice.

“The upstream DMR is methylated exclusively on the paternal allele, while the more downstream DMR is methylated exclusively on the maternal allele in both forward (129 x Cast) and reverse crossed (Cast x 129) hybrid mice.”

• cites [6] W Xie, 2012, Base-resolution analyses of sequence and parent-of-origin dependent DNA methylation in the mouse genome., Cell. sourceCould not verify

  The Results section on parent-of-origin-dependent methylation describes the 129 × Cast and Cast × 129 crosses, but the passages found do not identify an upstream and downstream DMR at Gnas/Nespas or their respective methylated alleles. Searches for Gnas, Nespas, and the stated DMR pattern found no passage establishing that locus-specific claim; page numbers were not available in our copy. (LLM, conf=0.72)

  Experimental — accuracy not yet measured

Probably not a substantial claim: may not pass S3 (load-bearing).

Spliced read aligners Tophat, STAR and HISAT2 have successfully been used for allele-specific alignments in conjunction with SNPsplit.

“Spliced read aligners that have successfully been used for allele-specific alignments in conjunction with SNPsplit include Tophat 18, STAR 14 (Spliced Transcripts Alignment to a Reference) and HISAT2 13.”

• cites [18] D Kim, 2013, TopHat2: accurate alignment of transcriptomes in the presence of insertions, deletions and gene fusions., Genome Biol. sourcePartly supported

  The Background describes TopHat2 as a spliced-read aligner, and the Results and discussion report successful spliced-read alignments. The cited paper does not establish that TopHat was used for allele-specific alignment in conjunction with SNPsplit; the citation attached to TopHat therefore supports the aligner identification but not that specific use. (LLM, conf=0.84)

  Experimental — accuracy not yet measured
• cites [14] A Dobin, 2013, STAR: ultrafast universal RNA-seq aligner., Bioinformatics. sourceNot checked — no full text of the source
• cites [13] D Kim, 2015, HISAT: a fast spliced aligner with low memory requirements., Nat Methods. sourcePartly supported

  The Results section (page not available) identifies HISAT as a spliced-read aligner and evaluates it alongside STAR and TopHat2. But the cited paper discusses HISAT and its two-pass variant HISATx2, not HISAT2, and our copy does not report allele-specific alignment with SNPsplit. The citation therefore supports the aligner description, not the claimed HISAT2–SNPsplit use. (LLM, conf=0.82)

  Experimental — accuracy not yet measured

May combine two claims.

In its generic mode SNPsplit is not tied to any particular aligner and operates across several different experiment types including RNA-Seq, genomic DNA-alignments, DNA methylation (Bisulfite-Seq) and 3-D genome organisation (Hi-C).

“In its generic mode SNPsplit is not tied to any particular aligner and operates across several different experiment types including RNA-Seq, genomic DNA-alignments, DNA methylation (Bisulfite-Seq) and 3-D genome organisation (Hi-C).”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

SNPsplit should perform well regardless of how the N-masking itself was accomplished, as long as a SAM/BAM file aligned to an N-masked genome is provided as input.

“As long as a SAM/BAM file that was aligned to an N-masked genome is provided as input SNPsplit should perform well regardless of how the N-masking itself was accomplished.”

May combine two claims.

N-masking the genome is a one-off exercise and requires only a single alignment to a reference, which noticeably reduces computational load.

“N-masking the genome is a one-off exercise and this approach has the advantage of requiring only a single alignment to a reference which noticeably reduces the computational load.”

May combine two claims.

N-masking effectively avoids allelic biases but may occasionally result in a minor loss of sensitivity when the density of N covered by a read is too high.

“Despite the fact that N-masking effectively avoids allelic biases it may occasionally result in a minor loss of sensitivity when the density of N covered by a read is getting too high.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

The SNPsplit genome preparation supports generation of single hybrid strains where one allele is the Black 6 reference and one alternative allele, and dual hybrid strains where both alleles are different from the Black 6 reference.

“The genome preparation step supports the generation of single hybrid strains where one allele is the same as the mouse reference sequence (which is based on strain C57BL/6J, hereafter called Black 6) and one alternative allele, e.g. SPRET/EiJ. It also supports the generation of dual hybrid strains where both alleles are different from the Black 6 reference, e.g. CAST/EiJ and 129S1/SvImJ.”

May combine two claims.

SNPsplit is not tied to any specific application and has been used already to answer questions for a variety of different data types such as ChIP-Seq, RNA-Seq, Bisulfite-Seq and Hi-C.

“SNPsplit is not tied to any specific application and indeed it has been used already to answer questions for a variety of different data types such as ChIP-Seq, RNA-Seq, Bisulfite-Seq and Hi-C.”

May combine two claims.

The majority of the genome shows very similar methylation levels on both alleles of the hybrid mice, but the approach allows detection of allele-specific methylation events.

“Whilst the majority of the genome shows very similar methylation levels on both alleles of the hybrid mice, this approach also allows the detection of allele-specific methylation events.”

Probably not a substantial claim: may not pass S3 (load-bearing).

SNPsplit operates in two stages: read tagging and read sorting.

“SNPsplit operates in two stages which are run sequentially: I) read tagging and II) read sorting.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Supported applications include genome re-sequencing, histone or protein ChIP-Seq or ATAC-Seq.

“A non-exhaustive list of supported applications includes genome re-sequencing, histone or protein ChIP-Seq (chromatin immunoprecipitation sequencing) or ATAC-Seq (Assay for Transposase-Accessible Chromatin by sequencing).”

Classified as probably not a substantial claim.

For hybrid mouse strains covered by the Mouse Genomes Project, SNPsplit offers an easy solution from generating N-masked genomes to allele-specific sorting of reads without requiring excessive computational skills.

“For hybrid mouse strains covered by the Mouse Genomes Project, SNPsplit offers an easy solution from generating N-masked genomes to allele-specific sorting of reads without requiring the user to possess excessive computational skills.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

For SNPs masked by Ns in the genome, no methylation call is performed during alignment; they receive a dot in the methylation call string, may be used for allele-sorting but do not participate in calling methylation.

“For SNPs which were masked by Ns in the genome no methylation call will have been performed during the alignment step, i.e. they will receive a ‘.’ (dot) in the methylation call string. This means that SNP positions themselves may be used for allele-sorting but do not participate in calling methylation.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

The automated genome preparation is optimised to work with the VCF file from the Mouse Genomes Project but can be adapted to work with any other genome as long as the genotypes are known and well defined.

“Even though the automated genome preparation is optimised to work with the VCF file from the Mouse Genomes Project the process can be easily adapted to work with any other genome as long as the genotypes are known and well defined.”

Classified as probably not a substantial claim.

SNPsplit is designed to enable researchers to quickly and easily perform allele-specific analysis of their sequencing data as long as the SNP genotypes of the organism in question are known.

“SNPsplit is designed to enable researchers to quickly and easily perform allele-specific analysis of their sequencing data as long as the SNP genotypes of the organism in question are known.”

Classified as probably not a substantial claim.

The SNPsplit genome preparation constructs N-masked genomes based on the Black 6 reference sequence using only high confidence homozygous SNP positions.

“The SNPsplit genome preparation first reads the SNP annotations for the strain in question from the VCF file and then constructs the N-masked genomes based on the Black 6 reference sequence using only high confidence homozygous positions.”

Probably not a substantial claim: may not pass S4 (based).

For SNPsplit to work, libraries must have been aligned to a genome in which all known SNP positions were masked with N and aligned using a suitable mapping program such as Bowtie2, TopHat, STAR, HISAT2, HiCUP or Bismark.

“For this to work libraries must have been aligned to a genome in which all known SNP positions were masked with the ambiguity base 'N' and aligned using a suitable mapping program such as Bowtie2, TopHat, STAR, HISAT2, HiCUP or Bismark.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Alignments to N-masked genomes require the aligner to support ambiguity DNA bases such as N.

“Alignments to N-masked genomes are not very different to regular mapping except that they require the aligner to support ambiguity DNA bases such as N.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

If the genotype is not known, SNP positions may be called from the data itself, or from genome re-sequencing performed in parallel.

“If the genotype is not known, SNP positions may be called from the data itself, or from genome re-sequencing performed in parallel.”

May combine two claims.

Masking SNP positions by Ns reduces slightly the number of total methylation calls but effectively eliminates the problem of assigning potentially incorrect methylation states to these positions.

“While this reduces slightly the number of total methylation calls it effectively eliminates the problem of assigning potentially incorrect methylation states to these positions.”

May combine two claims.

Positions requiring special treatment include C/T or T/C for forward strand alignments and G/A or A/G for reverse strand alignments; these positions may be used to assign opposing strand alignments because they do not involve C to T transitions directly.

“Positions requiring special treatment include all of the following Allele 1/Allele 2 combinations: C/T or T/C for forward strand alignments and G/A or A/G for reverse strand alignments. These positions may however be used to assign opposing strand alignments since they do not involve C to T transitions directly.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

To work smoothly with SNPsplit, HISAT2 and STAR require the user to disable soft-clipping, and STAR also needs to be instructed to print out the MD:Z: field.

“To work smoothly together with SNPsplit, HISAT2 and STAR require the user to disable soft-clipping which is performed by default (CIGAR operation: S), and STAR also needs to be instructed to print out the mismatch position (MD:Z:) field.”

May combine two claims.

SNPsplit requires reads to be a full match from end-to-end and does not support soft-clipping (CIGAR operation S).

“This means that SNPsplit requires reads to be a full match from end-to-end and thus soft-clipping (CIGAR operation: S), which may introduce artefactual alignments to poorly annotated regions in the genome 16 is not supported”

• cites [16] 2016, QC FailNot matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Probably not a substantial claim: may not pass S3 (load-bearing).

For mixed allele groups such as G1-G2 there is no need to create the reverse group G2-G1 because Hi-C interactions have no directionality.

“For mixed allele groups such as G1-G2 there is no need to create the reverse group (G2-G1) since Hi-C interactions have no directionality.”

• AssumptionRests on an assumption: Hi-C interactions have no directionality.