jgu的个人博客分享 http://blog.sciencenet.cn/u/jgu

博文

LRAcluster:基于低秩近似的多组学数据快速降维与聚类分析

已有 5375 次阅读 2015-7-5 19:27 |系统分类:论文交流

基于低秩近似的多组学数据快速降维与聚类分析

LRAcluster优点:能处理Gaussian、Possion、Binomial分布;速度快;保证低维空间正交

适用于大规模多组学数据的整合分析!欢迎使用!

http://bioinfo.au.tsinghua.edu.cn/software/lracluster 

=============================================

Fast dimension reduction and integrative clustering of multi-omics data usinglow-rank approximation: application to cancer molecular classification


Abstract

Background

One major goal of large-scale canceromics study is to identify molecular subtypes for more accurate cancerdiagnoses and treatments. To deal with high-dimensional cancer multi-omicsdata, a promising strategy is to find an effective low-dimensional subspace ofthe original data and then cluster cancer samples in the reduced subspace.However, due to data-type diversity and big data volume, few methods canintegrative and efficiently find the principal low-dimensional manifold of thehigh-dimensional cancer multi-omics data.

Results

In this study, we proposed a novellow-rank approximation based integrative probabilistic model to fast find the sharedprincipal subspace across multiple data types: the convexity of the low-rankregularized likelihood function of the probabilistic model ensures efficientand stable model fitting. Candidate molecular subtypes can be identified byunsupervised clustering hundreds of cancer samples in the reducedlow-dimensional subspace. On testing datasets, our method LRAcluster (low-rankapproximation based multi-omics data clustering) runs much faster with betterclustering performances than the existing method. Then, we applied LRAclusteron large-scale cancer multi-omics data from TCGA. The pan-cancer analysisresults show that the cancers of different tissue origins are generally groupedas independent clusters, except squamous-like carcinomas. While the singlecancer type analysis suggests that the omics data have different subtypingabilities for different cancer types.

Conclusions

LRAcluster is a very useful method forfast dimension reduction and unsupervised clustering of large-scale multi-omicsdata. LRAcluster is implemented in R and freely available via http://bioinfo.au.tsinghua.edu.cn/software/lracluster/.

 

Keywords

Mutli-omics, cancer, low-rankapproximation, clustering, dimension reduction, algorithm



附:艰难的审稿历程

1、1月初,先投ISMB2015,被拒

2、修改后,投Genome Biology、NAR,未外审

3、转投Bioinformatics,审稿后被拒,大约1个半月,初步判断3个Major

4、投PLoS Comput Biol,未外审


5、5月22日投BMC Genomics,到现在还在编辑部!"Editor Assigned"一个多月,两次询问编辑部后7月3日变成"Editor Invited",期间还给学术编辑发过一次信,没有任何回应;

7月21日状态又变成"Editor Assigned";

7月23日再次变成"Editor Invited"

7月29日变成"Editor Declined Invitation"

7月31日第三次变成"Editor Invited",当天又更新为"Editor Declined Invitation",然后又秒变成"Editor Invited"

8月5日变成第三次变成"Editor Declined Invitation"

8月6日再变成"Editor Assigned"

8月10日终于变成"Reviewers Assigned"

8月11日进入"Under Review"

9月7日第二次变成"Reviewers Assigned",几小时后变回"Under Review"

9月14日第一轮审稿终于完成"Reviews Completed"

9月18日给了"Major Revision",只有一个审稿人提了意见

9月29日论文修回,9月30日编辑部来信,说还有一个审稿人意见漏发了,稿子"Send back to authors",又得继续改!(真的是太奇葩了!)

10月8日论文修回

10月9日进入"Editor Assigned"

10月19日发信催了编辑部,10月21日终于把修改稿送审"Reviewers Assigned",几小时后进入"Under Review"

10月28日给了"Minor Revision",编辑说有个审稿人拖了很久没给意见,他自己根据回复给了决定

11月1日晚上11:30,第二次修回

11月2日进入"Editor Assigned"

11月5日开始"Editorial Assessment"

11月12日编辑部审核完毕要求补充两句伦理和作者知情的声明,当天修回

11月13日再次"Editorial Assessment"

11月16日终于Accepted



https://blog.sciencenet.cn/blog-407531-903070.html

上一篇:数据审编(data curation)——数据到知识的关键一步
下一篇:[转载]The Scientist: 肥胖与癌症 zz
收藏 IP: 111.196.82.*| 热度|

0

该博文允许注册用户评论 请点击登录 评论 (0 个评论)

数据加载中...
扫一扫,分享此博文

Archiver|手机版|科学网 ( 京ICP备07017567号-12 )

GMT+8, 2024-3-28 18:30

Powered by ScienceNet.cn

Copyright © 2007- 中国科学报社

返回顶部