进行很low的数据分析是一种什么体验？
2017-5-29 19:38
阅读：3534
标签：msap
分析某植物群体的分子数据， 想做一个高端大气的聚类树图，但使用的msap软件包只能绘制480*480像素的小图，样本一多，就特别难看，软件包没有直接提供可以调整的相关参数设置，而自己对R语言研究及其有限，于是我给软件的开发者发了一封邮件：
I want to use msap for drawing neighbor-Joining tree. It could only create picture with 480*480 pixels, it looks well with less than 30 samples. But I have about 90 samples , and I need creat a bigger picture. I don't know how to adjust the parameters.
很快收到了回信，给了一个链接，并说：
Anyway, take in mind that you can store the distance matrix and use it externally to plot different and improved trees (NJ tree in msap was intended as a raw exploratory analysis with very limited usefulness in most of the cases).
然后链接里解决方案的开头是这样写的：
The NJ plot form msap is awful, specially when the number of samples is big (>10). It was not intended to be very exahaustive, just as an exploratory analysis, but people seems to want to show it.
没错，我就是这个“people”，解决方案里的代码就是为我写，还是挺尴尬的，挺感动的。
这大牛可能搞不懂我这样的菜鸟为什么会有这么奇怪的需求， 这么low的需求，可我也明明看到有好几篇文献直接用了这个软件生成的NJ图，那几篇文章的水平还是挺不错。
菜鸟进阶之路不容易，一定要厚脸皮，相信以后如此low的叨扰会越来越少。
后记：解决了这个问题后。我居然又发现了问题，折腾了半天，发邮件询问，结果是我多输入一对引号而导致报错。囧囧囧。“Let me know if you have further problems/questions”，真是一个好的师者。
转载本文请联系原作者获取授权，同时请注明本文来自刘乐乐科学网博客。
链接地址：https://wap.sciencenet.cn/blog-1189133-1057836.html?mobile=1
收藏
当前推荐数：0
推荐到博客首页
网友评论0 条评论
