automatic update via todo-list
2024/08/10 (10.4h)
Doing tasks
Done tasks
- ‼edit my blog and organize files (10.4h)
Todo tasks in this week
- ‼模型描述: build the zero-evolution and add para; vim zhongqi_latex/src/03_model.tex and 00-daily*.md though
- learning model adding and config in revbayes
- make a model list for revbayes P3
- 输入文件格式，即数据来源
- 研究一下qmk，把轨迹球移动之后自动切换到第三层这件事给取消掉，再看看有什么可以配置玩的。
- align with magus？
- ON dating: run muscle in align and test mode via its online document.
- alignment is not much usefull on such huge dataset, why? or these gaps is normal on such taxa_num datasets? Filt them into small dataset will influent the pps_result?
- Can I get any acceptable pps_result?
- current running : mito_dataset. mft-ginsi/magus on dating; rb_pps on tower. check the result
Memo & Comments
- #char_type:335069
- edit blog in autodaily, add TODOLIST/WISHLIST in hexo, edit via obsidian.
- Write my thesis paper via mdbook;
- Maybe I should use mdbook instead of beamer for presentation, not hexo for documentation, maybe. #mdbook #hexo #todo #obsidian #blog
- Using todo list and hexo-blog via obsidian for daily work, using mdbook for presentation.
2024/08/10 (5.8h)
Doing tasks
Done tasks
- align with #magus ？ (0.0h) #alignment
- current running : mito_dataset. mft-ginsi/magus on dating; rb_pps on tower. check the result (2.3h) #alignment
- my vpn maybe broken, check it. (0.5h)
- config my workflow consist of todo/hexo/obsidian/mdbook/blog-README (3.0h)
Todo tasks in this week
- ‼模型描述: build the zero-evolution and add para; vim zhongqi_latex/src/03_model.tex and 00-daily*.md though
- learning model adding and config in revbayes
- make a model list for revbayes P3
- 输入文件格式，即数据来源
- 研究一下qmk，把轨迹球移动之后自动切换到第三层这件事给取消掉，再看看有什么可以配置玩的。
- ON dating: run muscle in align and test mode via its online document.
- alignment is not much usefull on such huge dataset, why? or these gaps is normal on such taxa_num datasets? Filt them into small dataset will influent the pps_result?
- Can I get any acceptable pps_result?
- Edit TODO-list and mdbook-files for ensuring they are in correct format.
Memo & Comments
- #char_type: 367368
- QiXi Festeval so that I did not much work today.
- Well, it’s weekend, just finish my workflow building tomorrow.
2024/08/12 (10.7h)
Doing tasks
Done tasks
- ‼ON dating: run muscle in align and test mode via its online document. (0.0h)
- ‼Edit TODO-list and mdbook-files for ensuring they are in correct format. (10.7h)
Todo tasks in this week
- build my init dataset_list #fit_data
- rewrite align.sh #fit_test
- clear my basic rule of my model. #robustness_model
Memo & Comments
#char_type : 403606;
align.sh was wrong, I should align codon, not just nuc. maybe all align should be rerun, but i just need choose only one. it is not a big thing now.
2024/08/13 (8.6h)
Doing tasks
Done tasks
- ‼build a init dataset_list, maybe just hemiptera but i should make them into datasets, not library. #fit_data (6.3h)
- get acc from shit_csv files. #fit_data (2.3h)
Todo tasks in this week
- rewrite align.sh #fit_test
- clear my basic rule of my model. #robustness_model
- ‼download data from acc #fit_data
Memo & Comments
#char_type : 434840;
2024/08/14 (8.1h)
Doing tasks
Done tasks
- ‼download data from acc (4.0h)
- align them into datasets. (4.1h)
Todo tasks in this week
- rewrite align.sh #fit_test
- clear my basic rule of my model. #robustness_model
- align mito_2012/2019 cds #fit_data
Memo & Comments
#char_type : 456049 ;
- mafft-xinsi for rrna and mafft-linsi for cds? #alignment
2024/08/15 (9.9h)
Doing tasks
- align mito_2012/2019 cds #fit_data (10.0h)
Done tasks
- rewrite align.sh #fit_test (1.0h)
- check linsi/ginsi/einsi diff. (0.7h)
- run GTR rb_pps_data in current dataset. #fit_test #pps (8.2h)
Todo tasks in this week
- ‼clear my basic rule of my model. #robustness_model
- current running: Dating: rb_pps. tower: mito_9/2_annotate; check the result.
Memo & Comments
#char_type : 486401;
- whole mito data cannot be split into genes simply. maybe I just need to test the whole align and the partial gene data. It’s not important so it will be ok.
- maybe I need to cut some column via seqkit or seqconverter because of gaps and maybe I can get better dataset.
2024/08/15 (8.7h)
Doing tasks
Done tasks
- ‼align mito_2012/2019 cds #fit_data (0.0h)
- current running: Dating: rb_pps. tower: mito_9/2_annotate; check the result. (2.7h)
- try to cut gap column after align(triml seqkit seqconverter). #fit_test #alignment (0.0h)
- ‼find the way of getting char/vars in rb. #fit_test #pps #revbayes (0.0h)
- find the way to use super computer. (0.0h)
- rb_pps for genomic dataset, bact, ahe (0.0h)
Todo tasks in this week
- ‼clear my basic rule of my model. #robustness_model
- understand the revbayes protocol
- current running: dating: bcod_rb_pps; tower: mito2019rna_align, bactCao_rbpps. check the result
- to be run: ahes_Cao_rbpps, mito2/9_rb_pps;
Memo & Comments
#char_type : 522966;
- list models, check every datasets. I need a table.
- Beside revbayes, I need some other tests, maybe.
- I need test the best model they used or best model under model-finder, not just GTR.
- Covarian and GHOST, how to test?
- maybe I should learn about how to enter model file into revbayes first, then think about my model.
2024/08/17 (11.0h)
Doing tasks
Done tasks
- current running: dating: bcod_rb_pps; tower: mito2019rna_align, bactCao_rbpps. check the result (1.4h)
- list models i should test and find out how to test/ input them into revbayes. #fit_test (9.6h)
Todo tasks in this week
- ‼clear my basic rule of my model. #robustness_model
- understand the revbayes protocol
- to be run: ahes_Cao_rbpps, mito2/9_rb_pps;
- I need a stand table for dataset/model/vars check.
Memo & Comments
#char_type : 541852
2024/08/18 (8.2h)
Doing tasks
Done tasks
- understand the revbayes protocol (0.0h)
- which covarian-like model i need to test? Should I discuss the “test conflict” between model fit and heterogenety test? #fit_test (5.5h)
- how to sample from pps? (2.7h)
Todo tasks in this week
- ‼clear my basic rule of my model. #robustness_model
- to be run: ahes_Cao_rbpps, mito2/9_rb_pps;
- I need a stand table for dataset/model/vars check.
- how to get dataset from treebase? #fit_data
Memo & Comments
#char_type : 570013
2024/08/19 (5.6h)
Doing tasks
Done tasks
- to be run: ahes_Cao_rbpps, mito2/9_rb_pps; (1.9h)
- I need a stand table for dataset/model/vars check. (0.0h)
- rewrite auto.sh (3.7h)
Todo tasks in this week
- clear my basic rule of my model. #robustness_model
- how to get dataset from treebase? #fit_data
- genrate genral pps protocol. not only revbayes. and make a change with the protocol in pp calculation (maybe parameter selection or model mixture) #fit_test
Memo & Comments
#char_type: 593248;
2024/08/20 (7.6h)
Doing tasks
Done tasks
- generate a couse viewer for datasets. #fit_cause (7.6h)
Todo tasks in this week
- clear my basic rule of my model. #robustness_model
- how to get dataset from treebase? #fit_data
- genrate genral pps protocol. not only revbayes. and make a change with the protocol in pp calculation (maybe parameter selection or model mixture) #fit_test
Memo & Comments
#char_type : 614842
- 今天发现之前的p-value计算脚本出错了，sim和emp的比较错了一位，已经改正
- 理解了R做图的脚本，发现没啥用，不如直接用pvalue脚本
- 理解了pvalue脚本的输出，其中low和upper就是emp值在pp分布中的上和下的概率，但是这都包括了恰好等于的部分，所以mid是把恰好等于取一半的修正，然后效应量是计算“模拟行为多大程度改变了样本总体”，效应量越小越好，太大不好。
2024/08/30 (24.5h)
Doing tasks
Done tasks
- how to get dataset from treebase? #fit_data (0.0h)
- genrate genral pps protocol. not only revbayes. and make a change with the protocol in pp calculation (maybe parameter selection or model mixture) #fit_test (18.7h)
- why there are some ‘nan’ in eff-test? (5.8h)
Todo tasks in this week
- clear my basic rule of my model. #robustness_model
- ‼generate phylobayes approach. #fit_test
Memo & Comments
#char_type : 666424
2024/08/31 (10.2h)
Doing tasks
Done tasks
- ‼generate phylobayes approach. #fit_test (10.2h)
- current run: pb_ances_test in tower and ahe_rb_pps in dating(because of memery limit in tower). (0.0h)
Todo tasks in this week
- clear my basic rule of my model. #robustness_model
Memo & Comments
#char_type : 696500
2024/09/04 (18.2h)
Doing tasks
Done tasks
- clear my basic rule of my model. #robustness_model (0.1h)
- ‼what is next? how long will i run? how to compare? (11.7h)
- learn covarion model and matrix in revbayes (2.9h)
- learn about Q matrix in markov process (3.5h)
Todo tasks in this week
- download data from treebase, generate a protocol. #fit_data
- ‼Simulate data in various rate/composition parameters. #fit_data
- PhyloMAd? Is it generate a different test? well, actually I just need to calculate the posterior predictive distribution for now. If PhyloMAd just generate different statistics, I can study it later.
Memo & Comments
#char_type : 769353
2024/09/05 (3.0h)
Doing tasks
- using parallel for pb_cm2. (-12.7h)
Done tasks
- learn about covarion model (3.0h)
Todo tasks in this week
- download data from treebase, generate a protocol. #fit_data
- ‼Simulate data in various rate/composition parameters. #fit_data
- PhyloMAd? Is it generate a different test? well, actually I just need to calculate the posterior predictive distribution for now. If PhyloMAd just generate different statistics, I can study it later.
Memo & Comments
#char_type : 826647
2024/09/08 (23.8h)
Doing tasks
Done tasks
- using parallel for pb_cm2. (0.0h)
- learn about covarion model (0.0h)
- check the cov model file (23.8h)
Todo tasks in this week
- download data from treebase, generate a protocol. #fit_data
- ‼Simulate data in various rate/composition parameters. #fit_data
- PhyloMAd? Is it generate a different test? well, actually I just need to calculate the posterior predictive distribution for now. If PhyloMAd just generate different statistics, I can study it later.
Memo & Comments
#char_type : 879369
2024/09/10 (13.7h)
Doing tasks
Done tasks
- build a mbl model file. (13.7h)
- 通过几个数据集计算phylobayes和revbayes的simulation是否有区别。后续全面转向phylobayes和其他语言吧，rev不靠谱。 (0.0h)
Todo tasks in this week
- download data from treebase, generate a protocol. #fit_data
- ‼Simulate data in various rate/composition parameters. #fit_data
- PhyloMAd? Is it generate a different test? well, actually I just need to calculate the posterior predictive distribution for now. If PhyloMAd just generate different statistics, I can study it later.
- 构建mmms和ghost的实现？或者找办法模拟一下。
Memo & Comments
#char_type : 907274
2024/09/11 (0.0h)
Doing tasks
Done tasks
Todo tasks in this week
- download data from treebase, generate a protocol. #fit_data
- ‼Simulate data in various rate/composition parameters. #fit_data
- PhyloMAd? Is it generate a different test? well, actually I just need to calculate the posterior predictive distribution for now. If PhyloMAd just generate different statistics, I can study it later.
- 构建mmms和ghost的实现？或者找办法模拟一下。
- JC_mbl的文件还差一点，就是var输出的时候格式有一点不对，pps跑不起来，应该是branchlenth输出不应该是向量的问题吧，test.var3可以跑，查看一下原因。
Memo & Comments
#char_type : 929101
2024/09/17 (0.0h)
Doing tasks
- ‼Simulate data in various rate/composition parameters. #fit_data (-19.8h)
Done tasks
- PhyloMAd? Is it generate a different test? well, actually I just need to calculate the posterior predictive distribution for now. If PhyloMAd just generate different statistics, I can study it later. (0.0h)
- 构建mmms和ghost的实现？或者找办法模拟一下。 (0.0h)
- JC_mbl的文件还差一点，就是var输出的时候格式有一点不对，pps跑不起来，应该是branchlenth输出不应该是向量的问题吧，test.var3可以跑，查看一下原因。 (0.0h)
Todo tasks in this week
- download data from treebase, generate a protocol. #fit_data
- read lie markov model paper.
Memo & Comments
#char_type : 968136
2024/09/19 (0.0h)
Doing tasks
- ‼Simulate data in various rate/composition parameters. #fit_data (-21.9h)
- read “alisim not good enough” paper (-22.4h)
Done tasks
- read lie markov model paper. (0.0h)
Todo tasks in this week
- download data from treebase, generate a protocol. #fit_data
Memo & Comments
#char_type : 983605
2024/09/22 (0.0h)
Doing tasks
- ‼Simulate data in various rate/composition parameters. #fit_data (-12.9h)
Done tasks
- read “alisim not good enough” paper (0.0h)
Todo tasks in this week
- download data from treebase, generate a protocol. #fit_data
- read fbd paper and learn about it.
Memo & Comments
#char_type : 992598
2024/09/24 (0.0h)
Doing tasks
Done tasks
Todo tasks in this week
- download data from treebase, generate a protocol. #fit_data
- Simulate data in various rate/composition parameters. #fit_data
- ‼read fbd paper and learn about it.
Memo & Comments
#char_type : 1020221
2024/09/27 (0.0h)
Doing tasks
- ‼read fbd paper and learn about it. (-13.9h)
Done tasks
Todo tasks in this week
- download data from treebase, generate a protocol. #fit_data
- Simulate data in various rate/composition parameters. #fit_data
Memo & Comments
#char_type : 1037327
2024/09/29 (0.0h)
Doing tasks
Done tasks
- ‼read fbd paper and learn about it. (0.0h)
Todo tasks in this week
- download data from treebase, generate a protocol. #fit_data
- Simulate data in various rate/composition parameters. #fit_data
- ‼learn about alisim in jc_cov or something(ghost?). #fit_data
Memo & Comments
#char_type : 1043569
2024/10/07 (0.0h)
Doing tasks
- ‼learn about alisim in jc_cov or something(ghost?). #fit_data (-13.9h)
Done tasks
Todo tasks in this week
- download data from treebase, generate a protocol. #fit_data
- Simulate data in various rate/composition parameters. #fit_data
- edit bin/simulator.sh in mac:workplace/test/. #fit_data
Memo & Comments
#char_type : 1059768
2024/10/14 (4.4h)
Doing tasks
Done tasks
- download data from treebase, generate a protocol. #fit_data (0.0h)
- Simulate data in various rate/composition parameters. #fit_data (0.0h)
- ‼learn about alisim in jc_cov or something(ghost?). #fit_data (4.4h)
- edit bin/simulator.sh in mac:workplace/test/. #fit_data (0.0h)
- test if alisim is same as revbayes pps. #fit_data (0.0h)
- covarion model format translate via phylogears #fit_data (0.0h)
Todo tasks in this week
- sync my simulator.sh into tower. #fit_data
Memo & Comments
#char_type : 1104641
2024/10/16 (1.7h)
Doing tasks
Done tasks
- sync my simulator.sh into tower. #fit_data (1.0h)
- zhongqii files. (0.0h)
- Liang Zonglei`s bpp. (0.7h)
- learn about <expected pps> article. (0.0h)
Todo tasks in this week
- R script learning #fit_cause
Memo & Comments
#char_type : 1115319
2024/10/31 (12.6h)
Doing tasks
Done tasks
- R script learning #fit_cause (3.5h)
- HPC learn. (7.7h)
- combine data and inf into one script. (1.4h)
- parallel auto in hpc. (0.0h)
- test mpi time used (0.0h)
- test parallel and mpi time used (0.0h)
Todo tasks in this week
- inferrence in R. #fit_cause
- learn about R analysis.
- error cascades pippeline.
- short seqs run under a suitable cores number.
- sims run experimental design.
Memo & Comments
#char_type : 1381138
2024/11/06 (0.0h)
Doing tasks
- build a donelist pipeline between hpc and tower (-13.6h)
Done tasks
- short seqs run under a suitable cores number. (0.0h)
- storage usage in hpc report; (0.0h)
- archive xz in auto.sh; (0.0h)
- done file list check between hpc and tower (0.0h)
- old done output. xz and tar. (0.0h)
Todo tasks in this week
- ‼inferrence in R. #fit_cause
- learn about R analysis.
- error cascades pippeline.
- ‼Based on F81 Model sets, design sims run experimental.
- read science causual infer paper.
- file translation;
Memo & Comments
#char_type : 1540811
2024/12/01 (0.0h)
Doing tasks
Done tasks
- ‼inferrence in R. #fit_cause (0.0h)
- learn about R analysis. (0.0h)
- read science causual infer paper. (0.0h)
- file translation; (0.0h)
- build a donelist pipeline between hpc and tower (0.0h)
- read causal inference. (0.0h)
- learn about tree comparation. rf distance; quartet distance. alias. (0.0h)
- ‼make a plot for zhongqii ppt. (0.0h)
- change the color of heatmap_smooth (0.0h)
- edit the code of raw_summary.txt.fmt in R. (0.0h)
- fig 4 add motation. (0.0h)
Todo tasks in this week
- error cascades pippeline.
- ‼Based on F81 Model sets, design sims run experimental.
- make a base matrix for analysis.
- my pps rev code. make topology prior be the same. test its influence in a small dataset.
- ‼read hohna paper 2023: model selection is no
- read yang ziheng paper : 2005 branch length prior influences.
- add convergence evaluation
Memo & Comments
#char_type : 1879514
2024/12/13 (0.0h)
Doing tasks
Done tasks
Todo tasks in this week
- error cascades pippeline.
- ‼Based on F81 Model sets, design sims run experimental.
- make a base matrix for analysis.
- my pps rev code. make topology prior be the same. test its influence in a small dataset.
- ‼read hohna paper 2023: model selection is no
- read yang ziheng paper : 2005 branch length prior influences.
- add convergence evaluation
- change ppt $problem to a frame.
- year summary. make a plan for next-step working.
Memo & Comments
#char_type : 2015115
2024/12/15 (0.0h)
Doing tasks
Done tasks
- make a base matrix for analysis. (0.0h)
- change ppt $problem to a frame. (0.0h)
- year summary. make a plan for next-step working. (0.0h)
- learn about argc (0.0h)
- make ai server env (0.0h)
Todo tasks in this week
- error cascades pippeline.
- ‼Based on F81 Model sets, design sims run experimental.
- my pps rev code. make topology prior be the same. test its influence in a small dataset.
- ‼read hohna paper 2023: model selection is no
- read yang ziheng paper : 2005 branch length prior influences.
- add convergence evaluation
Memo & Comments
#char_type : 2063913
2024/12/16 (0.0h)
Doing tasks
Done tasks
Todo tasks in this week
- “build your own group”. send a email to Prof. Luo Arong
- year-end summary slides for group willing.
- ‼integrate aichat and todo and my hexo blog together
Memo & Comments
#char_type : 2085460
#ai_shell:
你的工作包括文件和目录操作（如复制、移动）、配置文件编辑、脚本编写与调试、以及使用aichat等工具处理任务管理、信息查询等相关事务。
#aichat:
这两天的工作内容基本没有变化，包括向罗教授发送邮件、为意愿小组制作年终总结幻灯片以及将AI聊天工具、待办事项和Hexo博客整合在一起。
2024/12/18 (0.0h)
Doing tasks
Done tasks
- ‼integrate aichat and todo and my hexo blog together (0.0h)
- revbayes in rstudio and rmarkdown for beamer (0.0h)
Todo tasks in this week
- year-end summary slides for group will.
- statistics cor re-check
- clear the purpose that contacting Prof. Luo
- With the purposes, make a list that i need to do before sending email
Memo & Comments
#char_type : 2115970
#ai_shell:
你的工作涉及查看和处理Shell历史记录、编辑待办事项列表、使用aichat进行代码相关任务，以及运行和管理iqtree命令来分析基因序列。
此处iqtree主要是阅读文献PhyloForge，关于“SV Signal-Based Population Phylogeny”部分的复现。主要评价如下：
我翻了一下文中的代码，这里所说的sv系统发育，实际上是用indel的01矩阵跑iqtree和model finder…是以前处理形态矩阵时比较基本的一个处理，然而iqtree做树重建时没有考虑zipfian分布之类的indel模型，也不会因为二态矩阵而做相关处理，或许此处的创新可能在于对矩阵的编码方案与之前有所不同…然而依我拙见，sv在系统发育基因组的应用局限并不在重编码10矩阵上，否则可以通过“令xx为0，xx为1”的方式将任何事物变成“系统发育信号”…从这个角度上来讲，这篇文章提供的方案并不算是“基于sv的系统发育重建”，至少文中没有可见的论证…indel数据被应用到系统发育研究中已有很多解决方案，但似乎一直因有效性而难以推广，因而在基因组级别上并不常见，按此文说法或许这篇文章是第一次。数据扩大到基因组级后，有效性问题被数据淹没，增加采样频率确实可以降低信噪比，但考虑到相比系统发育研究的问题，样本总是寡而有偏的，这种信号的提高所带来的帮助是有局限和瓶颈的…况且，相比过采样，增加信号强度和更好的滤镜似乎对最终结果的影响是更大的。总之，注意到近年来系统发育基因组领域对sv数据的热情关注，我想良好的重建方法会在不久的将来出现。
#aichat:
这两天的工作变化集中在完成了将AI聊天、待办事项与Hexo博客集成的任务，同时在年终总结幻灯片和统计检查方面有所进展，并开始整理与教授罗联系的目的及其所需准备的事项。
2024/12/26 (0.0h)
Doing tasks
Done tasks
- download CLASSIC datasets for testing. (0.0h)
- use GTR alias as example, apply the Tame prior and moves. (0.0h)
Todo tasks in this week
- year-end summary slides for group will.
- statistics cor re-check
- ‼clear the purpose that contacting Prof. Luo
- With the purposes, make a list that i need to do before sending email
- read totally the Molecular Evolution by Ziheng Yang 2014.
- write a header comment block generator in rust
- the alpha in gamma 4 category prior: in a simple dataset, uniform(0,10^8) is better, but how about a heterogenity dataset? study on it.
Memo & Comments
#char_type : 2207749
今天仔细看了Tame先验和四足动物线粒体2024sb两篇文章的代码，以及revbayes的教程和源码文件，确认了先验、参数和moves的设置。使用gtr进行了标准化，其中branch lenth部分有比较大的更改，一些moves也有调整，应该重跑一些看看效果。
除此之外，branch rate可以在之后的time calibrating中设置，这是分区/全局的速率参数，用来计算绝对枝长的。目前我的分析应该用不到，但是以后其他研究中，不需要考虑是否必须在一次mcmc中全部采样的问题，branch lenth可以同时，也可以之后进行计算。
之后计算：https://revbayes.github.io/tutorials/sequential_bayes/stepwise_dating
同时计算：https://revbayes.github.io/tutorials/clocks/
2025/01/02 (0.0h)
Doing tasks
Done tasks
- year-end summary slides for group will. (0.0h)
- clear the purpose that contacting Prof. Luo (0.0h)
- With the purposes, make a list that i need to do before sending email (0.0h)
- read totally the Molecular Evolution by Ziheng Yang 2014. (0.0h)
- all models file. (0.0h)
- change the ‘model_name’ part. (0.0h)
Todo tasks in this week
- statistics cor re-check
- write a header comment block generator in rust
- the alpha in gamma 4 category prior: in a simple dataset, uniform(0,10^8) is better, but how about a heterogenity dataset? study on it.
- hpc/ dating/ tower : mcmc genertation, tsv2nex etc; change it in auto.sh, not one by one.
- heterotachy test in old papers. check them and calculate them.
Memo & Comments
#ai_shell:
#aichat: