w***g
发帖数: 5958 | 1 20 million row * 100 column, 即使存成text的话也就10G吧, 这种操作bottleneck在
磁盘, 假设100MB/s的话差不多读入2分钟, 写出2分钟, 无论如何10分钟应该够了啊.
你用numpy搞个大数组, 直接转置了输出难道有问题吗? |
|
w****w
发帖数: 521 | 2 你这个文件扫了快90次，当然慢了。根据内存多少要多列一起转置，而不是一列。你有
这么大内存，只要建个大矩阵，扫一次就可以了。而且你这里完全不用转置，直接产生
tped/tfam文件，而不是ped/map文件。 |
|
n******t
发帖数: 4406 | 3 这个帖子深刻的说明了，学语言如果只学这种所谓高级语言，是没法写任何能用的大型
系统的。 |
|
r****t
发帖数: 10904 | 4 reads like c code. why not directlly write in c. |
|
|
|
N******K
发帖数: 10202 | 7 这也是我为什要自己写matrix类的原因 因为可能要多线程等等
eigen类库有个搞笑的bug 还不好意思说是bug
a=a.transpose()； 这个结果就会出错 不知道怎么搞的 |
|
k****i
发帖数: 101 | 8 In haskell :)
ghci> :{
ghci| let fn' l1 l2 l3 =
ghci| map ( (x: y: z: _) -> x * y + z) .
ghci| Data.List.transpose $ [l1, l2, l3]
ghci| in fn' [1, 2, 3, 4] [2, 3, 4, 5] [4, 5, 6, 7]
ghci| ;}
[6,11,18,27] |
|
s******u
发帖数: 501 | 9 你可以去试一下，单机上跑fftw，MPI一点都不比OpenMP差，是不是更好我倒是忘记了
，很久以前跑的。有两个以上的node，比方说32x2或者32x4的话纯MPI就不行了，不过
这个也许跟MPI的实现和硬件有更大的关系。3d FFT的transpose要用到alltoall，如果
优化不好的话这个是最大的性能瓶颈
benchmark我找找看 |
|
|
w***g
发帖数: 5958 | 11 我已经给这作者发邮件去问了。我就觉得大家稀里哗啦发paper，里面水挺多的。
我也不是很确信，毕竟发到cvpr上去了，也是经过review的。 |
|
N*****m
发帖数: 42603 | 12 founder是老中啊。卫东你应该搞一个，估计也发了。 |
|
|
g****t
发帖数: 31659 | 14 现在这个行情，有真本事的，谁还写论文啊......
看文章都是浪费时间 |
|
q****p
发帖数: 4536 | 15 Thanks. I used paste special and check the Transpose box in excel. It did
not work in my way because my intention is not converting the rows into
columns or columns into rows in excel.
What I want is still the original table (say 5 rows and 10 columns), but in
a vertical display in word (still 5 rows and 10 cloumns). In other words you
have to flip over the table using 90 degree to the right. |
|
T*******n
发帖数: 493 | 16 # This gnuplot script will get you started.
# You need to transpose the data and save it in the format
# shown at the end of the script.
# If you want to color-fill the bars, you'll need to edit the
# EPS file that is created.
set terminal postscript eps enhanced color solid "Times-Roman" 20
set output "bargraph.eps"
set xrange [0:5]
set yrange [0:8000]
set xlabel "cases"
set ylabel "values"
set xtics ("case 1" 1, "case 2" 2, "case 3" 3, "case 4" 4)
set grid ytics
set mytics 5
set key width -9 b |
|
s*****t
发帖数: 1994 | 17 OK, what I used is exactly LSQR, and the problem what
I solve is exactly tomography.
Say, for one iteration, what I meant is one
A*x and then A^T * y, where ^T means transpose,
and y is the result from the previous iteration.
You know, if you want to speed up your code, it is
to speed up matrix multiplication A*x, and same way
A^T * y. The LSQR itself is not very time-consuming.
If you did not use LSQR, or other inversion code,
I believe your problem remains same, to speed up
matrix multiplicati |
|
t*******l
发帖数: 421 | 18 比如,我的数据如下:
1 3 4 6 7
3 5 2 4 4
3 3 5 6 4
怎么实现行列变换? |
|
q*****m
发帖数: 73 | 19 come on, it has nothing to do with un*x. |
|
t*******l
发帖数: 421 | 20 come on! it is!
有没有这样的命令,什么的??
unix! 干吗un*x??? |
|
|
g*********d
发帖数: 233 | 22 Box 1. Properties of euchromatic and heterochromatic regions
Trying to define heterochromatin is like trying to define life itself: a
cluster of important properties can be specified, but there are exceptions
in every instance. For example, centromeres are usually associated with
blocks of flanking heterochromatin. However, the inner centromere of
Drosophila chromosomes is associated with blocks of nucleosomes that contain
CENP-A (also known as CID), a variant of histone H3, interspersed with
bl... 阅读全帖 |
|
P****d
发帖数: 564 | 23 Why hard? :) It has been observed in the so called hybrid dysgenesis in
Drosophila and wallabies. In the Drosophila case, it involves the famous P
element, and has a fascinating story of co-evolution of human and Drosophila
and horizontal gene transfer at our own time http://engels.genetics.wisc.edu/Pelements/Pt.html.
I think that the key to McClintock's idea is not that transposable elements
can be activated, but that the activation is not merely due to damage of the
repressing machinery b... 阅读全帖 |
|
g******w
发帖数: 852 | 24 No1 ------------
Genomic imprinting in Drosophila has properties of both mammalian and insect imprinting.
Anaka M, Lynn A, McGinn P, Lloyd VK.
Dev Genes Evol. 2009 Feb;219(2):59-66. Epub 2008 Nov 25.
PMID: 19031081
[PubMed - indexed for MEDLINE]
No2---------------
Nat Rev Genet. 2007 Apr;8(4):272-85.
Transposable elements and the epigenetic regulation of the genome.
Slotkin RK, Martienssen R.
Source
Cold Spring Harbor Laboratory, 1 Bungtown Road, Cold Spring Harbor, New York 11724, USA.
... 阅读全帖 |
|
i**y
发帖数: 71 | 25 CpG island methylation is indeed highly repressive. But it doesn't happen
very often in normal cells. Most CpG island-methylated genes are germ-line-
specific genes or transposable elements.
The effect of sparse mCpG is still debated. There are also regions with
intermediate densities (e.g. CpG shores), where methylation anticorrelates
with gene expression. |
|
t*****t
发帖数: 773 | 26 Most CpG island-methylated genes are germ-line-
specific genes or transposable elements.
---------
这个信息很重要，谢谢！ 先前我倒没有注意到 |
|
|
q********i
发帖数: 4290 | 28 You may copy column E, righ click, select paste special,and select transpose. |
|
a*****x
发帖数: 901 | 29 你先在另一列（比如F）从开头（F2）写 "=E2*if（(row(E2)-INT(row(E2))/20=2,1,0)
"，然后拖一列。这样除了除以20余2的行都变成了0。然后你在filter里面把0都filter
掉，选全列copy/paste value到另一个worksheet,然后再copy/paste transpose成行。 |
|
a*****x
发帖数: 901 | 30 人家要贴等差间隔的格子。你这样整列都copy成行了。
transpose. |
|
a*****x
发帖数: 901 | 31 E2,E22,E42... --> 3P,3Q,3R...
不是说得很清楚了嘛。
我回答的也很清楚：
“1. 在另一列（比如F）从开头（F2）写 "=E2*if（(row(E2)-INT(row(E2))/20=2,1,0
)"，然后拖一列。这样除了除以20余2的行都变成了0；
2. 然后你在filter里面把0都filter掉；
3.选全列copy/paste value到另一个worksheet,然后再copy/paste transpose成行”
可惜没人给我发包子。。。 |
|
K****n
发帖数: 5970 | 32 所以才要先乘transpose矩阵，如果乘完了不可求逆就直接throw exception，线性回归
数据错误。
这些都是可以和面试官商量的
我也不喜欢求逆，避开矩阵直接求导就行了，也很trivial嘛 |
|
x********e
发帖数: 35261 | 33 自然突变的速度是恒定的，不过基因突变的主要途径不是mutation而是transposible
element。TE可不是恒速的。另外，任何变异都有冷却期，要稳定遗传突变需要的时间
不恒定 |
|
m******g
发帖数: 467 | 34 Science 13.11.15 (包含Editor's Choice等推荐的其他杂志的文章，之后省略这句话)
推荐：
1. Dosage compensation via transposable element mediated rewiring of a
regulatory network
D. miranda在1 my前出现neo-X chromosome, TE-mediated rewiring of regulatory
network (domestication, amplification, and fine-tuning) followed by erosion
of nonfunctional parts of the transposon
idea挺有趣的.
2. Long-term dynamics of adaptation in asexual populations
Power law model fits/predicts the data better than Hyperbolic model with an
asymptote.
Fitne... 阅读全帖 |
|
m******g
发帖数: 467 | 35 基本上就是题目+bulletin point的形式吧，不要纠结我对paper的品味，呵呵。新年快
乐！
PNAS 13.12.03
推荐：
1. Reintroducing domesticated wild mice to sociality induces adaptive
transgenerational effects on MUP expression
promiscuous / monogamous-line
higher MUP (pheromone) expression in pros-line
Male become more attractive
Difference in Mup promoter methylation level
2. High-throughput DNA sequencing errors are reduced by orders of magnitude
using circle sequencing
Previously ~0.1-1x10^-2
Unique library preparation strategy
Copie... 阅读全帖 |
|
m******g
发帖数: 467 | 36 Science 13.11.15 (包含Editor's Choice等推荐的其他杂志的文章，之后省略这句话)
推荐：
1. Dosage compensation via transposable element mediated rewiring of a
regulatory network
D. miranda在1 my前出现neo-X chromosome, TE-mediated rewiring of regulatory
network (domestication, amplification, and fine-tuning) followed by erosion
of nonfunctional parts of the transposon
idea挺有趣的.
2. Long-term dynamics of adaptation in asexual populations
Power law model fits/predicts the data better than Hyperbolic model with an
asymptote.
Fitne... 阅读全帖 |
|
m******g
发帖数: 467 | 37 基本上就是题目+bulletin point的形式吧，不要纠结我对paper的品味，呵呵。新年快
乐！
PNAS 13.12.03
推荐：
1. Reintroducing domesticated wild mice to sociality induces adaptive
transgenerational effects on MUP expression
promiscuous / monogamous-line
higher MUP (pheromone) expression in pros-line
Male become more attractive
Difference in Mup promoter methylation level
2. High-throughput DNA sequencing errors are reduced by orders of magnitude
using circle sequencing
Previously ~0.1-1x10^-2
Unique library preparation strategy
Copie... 阅读全帖 |
|
y**********n
发帖数: 478 | 38 bless宝宝
先尽力回答LZ在另外一个帖子里的前几个问题吧：
1 全基因测序是不是有很多遗传病其实也测不出来，是不是和数据库有关？
是的。只有确定了基因－疾病关系的才能检查出来。另外不同的数据库差别也比较大，
有的更新很慢。
刚才搜了下，这里有最新的关于NICCD的进展，估计新鉴定的这些突变都没有被分析。
By means of direct DNA sequencing, cDNA cloning and SNP analyses, 16 novel
pathogenic mutations, including 9 missense, 4 nonsense, 1 splice-site, 1
deletion and 1 large transposal insertion IVS4ins6kb (GenBank accession
number KF425758), were identified in CTLN2 or NICCD patients from China,
Japan and Malaysia, respectively, making the SLC25A1... 阅读全帖 |
|
z****u
发帖数: 1007 | 39 不太明白这点，piggybac过表达难道不是引起transpose跳转吗？不明白为啥会把俩itr
间的序列切除掉？
变。 |
|
h****b
发帖数: 10 | 40 来自主题: Computation版 - 对矩阵求导 查了很多数学书、数学手册，都没有找到详细的有关对矩阵求导的，从internet
上找到了一些只言片语。比如：
（请注意这里的 Y A 和 X 都是矩阵，A * X 表示标准矩阵乘法，由于这里没法
表示复杂的数学符号和上下标等，我用 A' 表示 A 的转置阵 transpose(A),
DY/DX 表示 Y 对 X 求导）
Y = A * X --> DY/DX = A'
Y = X * A --> DY/DX = A
Y = A' * X * B --> DY/DX = A * B'
Y = A' * X' * B --> DY/DX = B * A'
我做的一些矩阵运算中还包括了矩阵的Hadamard product(元素与元素相乘)，
如果用 A 。X 表示 Hadamard product，类似于上面那些式子，Y 对 X 求导
应该怎么做呢？比如：
Y = A 。X --> DY/DX = ？
Y = X 。A --> DY/DX = ？
Y = A' 。X 。B --> DY/DX = ？
Y = A' 。X' 。B --> |
|
e***n
发帖数: 286 | 41 Urgent!
For any symmetric definite matrix, for sure we can factor it with Cholesky
method. How about symmetric indefinite matrix? I need factor such a matrix
A exactly into the product form
A = B * B'
where B' is the transpose of B and B is some n x n matrix ( not necessarily
to be triangular).
I know we can factor it with a LDLT method and further get
A = P * P' - Q * Q'
But this is not exactly what I need.
Thank you very much! |
|
c*******a
发帖数: 18 | 42 I think that can be solved in this way.
let A be your equation matrix
X be your parameters, B is your result vector
so AX = B
Where A is a 10,000x100 matrix, X is a 100x1 matrix
B is a 10,000x1 matrix
we need to solve x, so we multiply A'(A transpose to both side)
we get (A'A)X=A'B
Now A'A becomes a 100x100 square matrix,so we can calculate the inverse
multiply the inverse on both side
we get X=(A'A)^(-1) A'B
That is just a illustration of the method, there should be more
efficient method to ac |
|
c****r
发帖数: 576 | 43 【 以下文字转载自 Mathematics 讨论区 】
发信人: celler (生活正常化了), 信区: Mathematics
标 题: matlab 能不能解符号矩阵方程？
发信站: BBS 未名空间站 (Mon Nov 26 17:34:14 2007)
这样的方程：J*C + C*J' = D,其中J' = transpose(J)，J，C，D都是带有符号变量的
矩阵，求符号解。 |
|
h***r
发帖数: 726 | 44 【 以下文字转载自 Mathematics 讨论区 】
发信人: haier (no nickname), 信区: Mathematics
标 题: karmarkar's paper
发信站: BBS 未名空间站 (Wed Mar 3 17:00:58 2010, 美东)
after the ellipsoid paper, Karmarkar developed the interior point method for
LP.
Does anyone understand why the canonical form, on the first page, is the
following? (WHY Z, not R?)
min x' x
s.t. Ax = 0
\sum xi = 1
x >= 0
x \in R^n, c \in Z^n, A \in Z^(mn)
x' means x transpose
R^n means R to the nth
So, Why c, A in Z, not R???
Thanks! |
|
b***k
发帖数: 2673 | 45 【 以下文字转载自 Mathematics 讨论区 】
发信人: blook (布鲁克), 信区: Mathematics
标 题: 线性代数一问
发信站: BBS 未名空间站 (Sun Mar 16 23:42:21 2014, 美东)
Given an arbitrary matrix A with dimension m x n,
what is the difference or relationship of A'*A and A*A'?
here A' is transpose of A.
obviousely both of them are symmetric, and they have identical eigenvalues.
but their eigenvectors are different which means eigenspace of both matrix
are distinct, right? |
|
w*****1
发帖数: 473 | 46 【 以下文字转载自 Statistics 讨论区 】
发信人: wz99331 (dotti), 信区: Statistics
标 题: ask for help for R programming
发信站: BBS 未名空间站 (Thu Oct 16 15:25:16 2014, 美东)
I need to transpose a file to example.txt
the final dataset should be sorted by chr first then by position. Could
anyone help me to write the R code? Thank you very much!
raw data:
snp Chr pos Sample.ID Allele1 Allele2
rs6269442 1 1.614 517 A B
rs13475701 1 1.744 517 ... 阅读全帖 |
|
s*********n
发帖数: 149 | 47 First, I don't think this is a good idea to combine multiple record to one;
but if you insist to do so, you can use array or proc transpose;
for Array,
first you have to find out the max of each id has. so that you can define you
array : array age(*) $1. age1-age20;
also you need to define array for each variable.
say the data set call temp, sort by id;
data temp;
set temp;
by id;
if first.id then
dimnum = 1;
age(dimnum) = age;
else dimnum ++1;
if last.id then output;
run;
also you can creat ano |
|
f******k
发帖数: 297 | 48 given a square matrix A, what is the minimum of
abs((x'Ax)/(|Ax||x|)) over all nonzero x, where |.| is
Euclidean norm and x' is the transpose of x? Or
equivalently what is the minimum angle between Ax and x? |
|
p*****w
发帖数: 429 | 49 很久没接触距阵了
求||W||的值
X= [a,b]
W= transpose of X |
|
b*****e
发帖数: 1 | 50 By definition, rotations preserve length and origin, which means
for all points x and y in the space. |T(x)-T(y)|=|x-y|, where T is
the rotation, and T(0)=0. Easy to see such a rotation is a
linear operation.
Let ' donate transpose of vectors and matrices.
● is a dot product.
Since |T(x)-T(y)|=(T(x)-T(y●(T(x)-T(y))=(x-y) ●x-y), let y=0,
we can get T(x) ●T(x)=x●x. Also, since T is linear, we can
write T(x) as Ax, where A is a n by n matrix. Then we have:
T(x) ●T(x)= T(x)'T(x) =(Ax)'(Ax)=x'A'Ax= |
|
