In this figure we show the output of the half-loop biclustering algorithm applied to the BDRN cohort in arm-1 (limited to those SNPs with maf ≥0.25).

<p>As described in the main text, the algorithm proceeds iteratively, eliminating rows and columns from the case-subject-array <i>D</i> until all have been removed. At each iteration <i>i</i>, the remaining submatrix <i>D</i>(<i>i</i>) comprises...

Full description

Saved in:
Bibliographic Details
Main Author: Caroline C. McGrouther (5216153) (author)
Other Authors: Aaditya V. Rangan (14659986) (author), Arianna Di Florio (20636356) (author), Jeremy A. Elman (9602147) (author), Nicholas J. Schork (176346) (author), John Kelsoe (3437189) (author)
Published: 2025
Subjects:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1852023186431934464
author Caroline C. McGrouther (5216153)
author2 Aaditya V. Rangan (14659986)
Arianna Di Florio (20636356)
Jeremy A. Elman (9602147)
Nicholas J. Schork (176346)
John Kelsoe (3437189)
author2_role author
author
author
author
author
author_facet Caroline C. McGrouther (5216153)
Aaditya V. Rangan (14659986)
Arianna Di Florio (20636356)
Jeremy A. Elman (9602147)
Nicholas J. Schork (176346)
John Kelsoe (3437189)
author_role author
dc.creator.none.fl_str_mv Caroline C. McGrouther (5216153)
Aaditya V. Rangan (14659986)
Arianna Di Florio (20636356)
Jeremy A. Elman (9602147)
Nicholas J. Schork (176346)
John Kelsoe (3437189)
dc.date.none.fl_str_mv 2025-01-29T18:23:47Z
dc.identifier.none.fl_str_mv 10.1371/journal.pone.0314288.g002
dc.relation.none.fl_str_mv https://figshare.com/articles/figure/In_this_figure_we_show_the_output_of_the_half-loop_biclustering_algorithm_applied_to_the_BDRN_cohort_in_arm-1_limited_to_those_SNPs_with_maf_0_25_/28304632
dc.rights.none.fl_str_mv CC BY 4.0
info:eu-repo/semantics/openAccess
dc.subject.none.fl_str_mv Genetics
Neuroscience
Pharmacology
Biotechnology
Evolutionary Biology
Marine Biology
Mental Health
Infectious Diseases
Plant Biology
Biological Sciences not elsewhere classified
termed &# 8216
psychiatric genomics consortium
leverage recent advances
e ., biclusters
corrected biclustering algorithm
subgroup may represent
subgroup &# 8217
pgc containing 5781
various bd studies
bd subtype information
improve risk prediction
specific pattern highlighted
homogeneous genetic subgroup
relatively small subset
genetically homogeneous subtype
discovered without using
bicluster &# 8217
specific genetic pattern
correlated genetic risk
2524 bd cases
genetic heterogeneity delineating
disorder </ p
specific pattern
genetic subgroup
bd cases
genetic level
genetic data
pgc ).
9752 cases
bipolar disorder
xlink ">
successfully identified
subphenotypic data
subjects define
set aside
remaining data
related signal
prs ),
primary focus
particularly notable
one way
informed polygenic
gwas analyses
fully understood
first apply
expression across
eliminate noise
driven subgroups
challenging task
broad phenotype
dc.title.none.fl_str_mv In this figure we show the output of the half-loop biclustering algorithm applied to the BDRN cohort in arm-1 (limited to those SNPs with maf ≥0.25).
dc.type.none.fl_str_mv Image
Figure
info:eu-repo/semantics/publishedVersion
image
description <p>As described in the main text, the algorithm proceeds iteratively, eliminating rows and columns from the case-subject-array <i>D</i> until all have been removed. At each iteration <i>i</i>, the remaining submatrix <i>D</i>(<i>i</i>) comprises case-subjects and allele-combinations . At each iteration we record the ‘row-trace’ , which is the covariate-corrected average level of differential-expression between <i>D</i>(<i>i</i>) and the control-subjects <i>X</i>. In the top row of subplots we show the row-trace for the data (red) as well as for 128 label-shuffled trials (black). Each of the row-traces has been transformed into an iteration-dependent z-score (estimated using the distribution of label-shuffled trials at that iteration). In the bottom row we show the corresponding empirical p-value, as estimated for each iteration using the label-shuffled trials. The dashed black-line corresponds to the 95th percentile (i.e., a significance value of 0.05 if each iteration were considered independently). If the signal were homogeneous we would expect to see the red trace begin at a high value and decay relatively monotonically. By contrast, we see strong evidence for heterogeneity; the red trace is far from monotonic. The overall p-value for the data (red-trace), estimated using the strategy in [<a href="http://www.plosone.org/article/info:doi/10.1371/journal.pone.0314288#pone.0314288.ref064" target="_blank">64</a>], is <i>p</i> ≲ 1/64. Note that the trace is significant over a range of iterations, including <i>i</i> ∈ [175, 350].</p>
eu_rights_str_mv openAccess
id Manara_2aff7a3bc66980a01cec0b6b8e26f2ef
identifier_str_mv 10.1371/journal.pone.0314288.g002
network_acronym_str Manara
network_name_str ManaraRepo
oai_identifier_str oai:figshare.com:article/28304632
publishDate 2025
repository.mail.fl_str_mv
repository.name.fl_str_mv
repository_id_str
rights_invalid_str_mv CC BY 4.0
spelling In this figure we show the output of the half-loop biclustering algorithm applied to the BDRN cohort in arm-1 (limited to those SNPs with maf ≥0.25).Caroline C. McGrouther (5216153)Aaditya V. Rangan (14659986)Arianna Di Florio (20636356)Jeremy A. Elman (9602147)Nicholas J. Schork (176346)John Kelsoe (3437189)GeneticsNeurosciencePharmacologyBiotechnologyEvolutionary BiologyMarine BiologyMental HealthInfectious DiseasesPlant BiologyBiological Sciences not elsewhere classifiedtermed &# 8216psychiatric genomics consortiumleverage recent advancese ., biclusterscorrected biclustering algorithmsubgroup may representsubgroup &# 8217pgc containing 5781various bd studiesbd subtype informationimprove risk predictionspecific pattern highlightedhomogeneous genetic subgrouprelatively small subsetgenetically homogeneous subtypediscovered without usingbicluster &# 8217specific genetic patterncorrelated genetic risk2524 bd casesgenetic heterogeneity delineatingdisorder </ pspecific patterngenetic subgroupbd casesgenetic levelgenetic datapgc ).9752 casesbipolar disorderxlink ">successfully identifiedsubphenotypic datasubjects defineset asideremaining datarelated signalprs ),primary focusparticularly notableone wayinformed polygenicgwas analysesfully understoodfirst applyexpression acrosseliminate noisedriven subgroupschallenging taskbroad phenotype<p>As described in the main text, the algorithm proceeds iteratively, eliminating rows and columns from the case-subject-array <i>D</i> until all have been removed. At each iteration <i>i</i>, the remaining submatrix <i>D</i>(<i>i</i>) comprises case-subjects and allele-combinations . At each iteration we record the ‘row-trace’ , which is the covariate-corrected average level of differential-expression between <i>D</i>(<i>i</i>) and the control-subjects <i>X</i>. In the top row of subplots we show the row-trace for the data (red) as well as for 128 label-shuffled trials (black). Each of the row-traces has been transformed into an iteration-dependent z-score (estimated using the distribution of label-shuffled trials at that iteration). In the bottom row we show the corresponding empirical p-value, as estimated for each iteration using the label-shuffled trials. The dashed black-line corresponds to the 95th percentile (i.e., a significance value of 0.05 if each iteration were considered independently). If the signal were homogeneous we would expect to see the red trace begin at a high value and decay relatively monotonically. By contrast, we see strong evidence for heterogeneity; the red trace is far from monotonic. The overall p-value for the data (red-trace), estimated using the strategy in [<a href="http://www.plosone.org/article/info:doi/10.1371/journal.pone.0314288#pone.0314288.ref064" target="_blank">64</a>], is <i>p</i> ≲ 1/64. Note that the trace is significant over a range of iterations, including <i>i</i> ∈ [175, 350].</p>2025-01-29T18:23:47ZImageFigureinfo:eu-repo/semantics/publishedVersionimage10.1371/journal.pone.0314288.g002https://figshare.com/articles/figure/In_this_figure_we_show_the_output_of_the_half-loop_biclustering_algorithm_applied_to_the_BDRN_cohort_in_arm-1_limited_to_those_SNPs_with_maf_0_25_/28304632CC BY 4.0info:eu-repo/semantics/openAccessoai:figshare.com:article/283046322025-01-29T18:23:47Z
spellingShingle In this figure we show the output of the half-loop biclustering algorithm applied to the BDRN cohort in arm-1 (limited to those SNPs with maf ≥0.25).
Caroline C. McGrouther (5216153)
Genetics
Neuroscience
Pharmacology
Biotechnology
Evolutionary Biology
Marine Biology
Mental Health
Infectious Diseases
Plant Biology
Biological Sciences not elsewhere classified
termed &# 8216
psychiatric genomics consortium
leverage recent advances
e ., biclusters
corrected biclustering algorithm
subgroup may represent
subgroup &# 8217
pgc containing 5781
various bd studies
bd subtype information
improve risk prediction
specific pattern highlighted
homogeneous genetic subgroup
relatively small subset
genetically homogeneous subtype
discovered without using
bicluster &# 8217
specific genetic pattern
correlated genetic risk
2524 bd cases
genetic heterogeneity delineating
disorder </ p
specific pattern
genetic subgroup
bd cases
genetic level
genetic data
pgc ).
9752 cases
bipolar disorder
xlink ">
successfully identified
subphenotypic data
subjects define
set aside
remaining data
related signal
prs ),
primary focus
particularly notable
one way
informed polygenic
gwas analyses
fully understood
first apply
expression across
eliminate noise
driven subgroups
challenging task
broad phenotype
status_str publishedVersion
title In this figure we show the output of the half-loop biclustering algorithm applied to the BDRN cohort in arm-1 (limited to those SNPs with maf ≥0.25).
title_full In this figure we show the output of the half-loop biclustering algorithm applied to the BDRN cohort in arm-1 (limited to those SNPs with maf ≥0.25).
title_fullStr In this figure we show the output of the half-loop biclustering algorithm applied to the BDRN cohort in arm-1 (limited to those SNPs with maf ≥0.25).
title_full_unstemmed In this figure we show the output of the half-loop biclustering algorithm applied to the BDRN cohort in arm-1 (limited to those SNPs with maf ≥0.25).
title_short In this figure we show the output of the half-loop biclustering algorithm applied to the BDRN cohort in arm-1 (limited to those SNPs with maf ≥0.25).
title_sort In this figure we show the output of the half-loop biclustering algorithm applied to the BDRN cohort in arm-1 (limited to those SNPs with maf ≥0.25).
topic Genetics
Neuroscience
Pharmacology
Biotechnology
Evolutionary Biology
Marine Biology
Mental Health
Infectious Diseases
Plant Biology
Biological Sciences not elsewhere classified
termed &# 8216
psychiatric genomics consortium
leverage recent advances
e ., biclusters
corrected biclustering algorithm
subgroup may represent
subgroup &# 8217
pgc containing 5781
various bd studies
bd subtype information
improve risk prediction
specific pattern highlighted
homogeneous genetic subgroup
relatively small subset
genetically homogeneous subtype
discovered without using
bicluster &# 8217
specific genetic pattern
correlated genetic risk
2524 bd cases
genetic heterogeneity delineating
disorder </ p
specific pattern
genetic subgroup
bd cases
genetic level
genetic data
pgc ).
9752 cases
bipolar disorder
xlink ">
successfully identified
subphenotypic data
subjects define
set aside
remaining data
related signal
prs ),
primary focus
particularly notable
one way
informed polygenic
gwas analyses
fully understood
first apply
expression across
eliminate noise
driven subgroups
challenging task
broad phenotype