Getting Netbook SD card reader to work in Ubuntu

Had installed the new Ubuntu 11.04 on an Aspire One but could not get the card to work. Googling did not help. Knew that the Aspire (and a lot of netbooks) had a Realtek reader. Went to the Realtek website to see if they had a linux driver. Yes they did! Downloaded and installed as below and got the SD card reader to work.
Instructions.
Step 1: Download the Realtek driver (link)
Step 2: Drag the downloaded file into your home folder
Step 3: Right click on the file and select "Extract here"
Step 4: Open up the terminal.
Step 5: Type in "cd rts_pstor" (without quotes). Press Return. If any error message appears, make sure that the realtek driver file has been extracted in your home directory (the one with your user name) and that what you type in after cd is the same as the extracted file folder.
Step 6: Wait till the prompt come ups. Then type in "make". Press Return.
Step 7: Then type in "sudo make install". Type in your admin password when prompted.
Step 8: Type in "sudo depmod". Return.
Step 9: Restart the netbook. You should now see yr SD card when inserted.

Notes from Beyond genome meeting, Sept 21 2011, Rockville

Jay Shendure, Univ Washington. 2 yrs ago array hybzn based capture of 12 exomes for Mendelian disorders (ng et al 2009). Now soln based 96 plex exome capture, 4k sequenced by Univ Washington. 100,000’s of genomes in 2-3 years the way prices are dropping. Kabuki syndrome, gene discovery, multiple malformation syndrome, suspected dominant, 10 probands sequenced, found MUC16, thought artifact, failure to identify … dialed back our analysis combined with phenotypic stratification … now came up with MLL2 (histone methyltransferase), all were de novo mutations (so could not be solved by mapping). 110 cases sequenced for MLL2 by Sanger (“hard to find Sanger sequencers”). 74% with mutation, loss of function (usually truncating), almost always de novo. 24% are mutation negative for MLL2 though phenotypically indistinguishable from mutation positive cases. So sequenced the entire the 50kb MLL2 locus, used transposase based library prep, 48plex capture on array. Did not find any additional changes in the MLL2 locus. So a second gene? Also for MLL2 which mutations are pathogenic and which are benign. Autism Spectrum Disorders, complex phenotypes and genetics. Started with trio sequencing on 209 trios (627 exomes), expect 1 de novo in coding per generation. Used NimblegenEZ Exome V2.0 (29.5 Mb), 90% on target. Filter out common variants, sanger the final ones. Found 242 new de novo variants, confirmed (mutation rate of 2.17 x 10-8 per generation). 2 cases parental and 9 cases somatic mosaicism. 4:1 bias in nonsense mutation rate when compared to private inherited mutations from 1776 exomes (ESP ”controls”). 125 mutations classified as severe (truncating, missense (Grantham score). 40% of mutations part of 1 network. Individuals with multiple hits, e.g, 1 individual with 5 mutations, some inherited some de novo. Extreme genetic heterogeneity but only 2 recurrently disrupted genes. How to validate candidate genes? Used molecular inversion probes (circular) for 6 genes with 384-plex per HiSeq lane. Performed on 2400 individuals. $10.65 per sample was the cost for capture and sequencing. General notes: Does not always work to identify the causative gene. Mouse model with very well known genetics indicating a locus but could not find the causative gene.

Tim O’Connor, U Wash Genome Sci, 2440 human exomes, ESP, popn genetics. NHLBI Exome Sequencing Project (ESP) by Broad's David Altshuler, and Debbie Nickerson from Seattle, goal to do about 250000 samples. How much variation in coding regions in populations,. Started with 2500 exomes, mapped to hg19, yielded 1.2M SNV from 114x coverage. Filtering produced 56k SNVs from 22.4 Mb region. 2 ethnicities, euro Americans and African Americans, 91k SNVs shared. 82% of SNVs novel.  6k nonsense mutations. Slides on types of mutations in the exomes, synonymous to indels. Number of SNVs as a function of sample size (common vs rare variants). Nucleotide diversity across genes and across genes organized by pathways. HLA and immune relared pathways very high, house keeping very low. 14k variants per person, about 600 would be deleterious. Ancestry,

Ron Do, Harvard Med Sch. Exome myocardial infarction, LDL cholesterol,  GWAS for coronary disease up to 30 loci, what about rare variants? type 1 by APOB mutation, type 2, non-APOB. Siblings with most extreme (low) lipid values, exome sequencing at Broad, 200x covg., ILMN GAII, GATK analysis. 18k unfiltered varoiants, removing dbSNP leaving 746 SNVs, 1kGP and ESP variants removed, leaving 1 gene shred in both siblings, ANGPTL3 nonsense variants. IGV viewer. Sanger validation. 38 available family members, carriers with mutations, siblings with both mutations. 10 children from one couple which “helped our study”. Expanded to other families, confirmed. The GWAS on 100k individuals had a rare allele concordant with this study.  Myocardial infarct study, unrelated individuals with extreme phenotypes. NHLBI exome sequencing project, large collaboration group, 1200 cases (young MI), 1200 controls (old without MI). Currently done 500+ 500 exomes. 177x covg., 16k known SNPs, association tests for rare variants (Li & Leal, 2008). No strong associations found, though picked NBEAL1 as GWAS picked locus. Limited GWAS with 200 SNPs on 61k samples, rediscovered PCSK9c679 gene protective to MI, African-american specific. Through imputation discovered CDKN2BAS, LPA, MIA3, ALS2CR8, WDR12, LIPA, … many reported in previous GWAS. So what sample size for exome to reach significance, p-value increases substantially from  500+ samples to 5000+ samples (6x10-14). So 5000+ samples for a definitive result for complex traits. Good talk for synergistic use of genotyping and sequencing.

Joris Veltman, Nijmegen, a de novo paradigm, intellectual disability. De novo mutations versus inherited diseases. 100-500 de novo mutations per generation. Mutational target size determines frequency of disease. Sporadic occurrences of diseases. Linkage and association studies limited in some cases. Intellectual disability, 2% of population, heritability of 0.8-0.9. 5% chromosomal abnormality not in parents. De novo CNVs, offered as a diagnostic. E.g., Deletion of 600 Kb at 17q21.31 in kids resulting from inversion in same region from parents. Showed 5500xl in a slide, trio-aapproach to filter for de novo mutations. Pilot study on 10 patient-parent trios. No de novo CNV. Agilent exome. 5 variants after filtering. 0-2 validated by sanger. 10 patients, 9 different genes, found 2 previously reported genes (RAB38B, SYNGAP1) involved in mental retardation. Now on 600 patients, sequencing of YY1, transcription factor involved in memory and plasticity. 40-50% of patients had a de novo mutation, plus additional 10-15% de novo CNV.

Gholson Lyon, Childrens Hospital, Philadelphia. Tolstoy quote, happy families alike, unhappy families all different in different ways. Moved to Univ Utah, started Utah foundation for family studies. Helena Black with kids and grand kids who had the little old man phenotype. Unknown syndrome, x-chromo capture and sequencing, agilent and illumina. hNaa10p-S37P. Protein acetylation.

Stephen Kingsmore, Childrens Mercy hosp, Kansas City. Clinical seq apps, 7k mendelian disorders as of today, 3k molecula basis known, disease burden 4x higher in more inbred populations. Monogenic diseases inheritance, autosomal, x-linked, dominant recessive, mtDNA, imprinting, etc. About 1k diseases ready for prime time diagnostic tests but only 200-400 disease gene tests available. Current 50% success rate by Sanger methods at a cost of about $10k/patient. Newborn screening for 120 disorders. Preconception carrier testing for 100 disorders. Fetal tests. Pharmacogenomics. 2 ways for diag test, thru FDA approved test or thry CLIA approved lab. Not enough volume for fda tests. Cost of DNA  sequencing dropped 5000 fold in the last few years. Gene patents ruled “unsustainable”. Mendelian genomic medicine. NGS mol dx is about $1/disease/patient compared to $10k/disease/patient. 619 diseases x 96 patients at once by NGS. Exome too expensive, misses exonic and GC rich regions. As soon as price comes down we would move to whole genome sequencing. Currently, 0.07% of the genome enriched, exons/introns relevant to 619 diseases, 4 Gb sequencing, interpretation. Requires about $3M, 1/3^rdmillion for compute. Favor ILMN targeted approach, beats the competition, step temperature, better. ILMN hiseq v3, 1x101 bp/5 days. How deep do u need to sequence? Stat slide on truseq capture efficiency. Beats array technology in performance, sensitivity, specificity. Example. 4 brothers in whom conventional testing failed. Sequenced brother 3, found CLEC7A, known pathological mutation also another one DOCK8, unknown, … if they had the result they would have been offered hematopoeitic stem cell transplants. No clinical grade database. NCBI – new database called ClinVar. LOVD in Europe.

Notes from Beyond Genome Sept 20 2011 meeting in Rockville

DNAstar vendor presentation. Good desktop performance, alignment of human genome in 24 hrs, de novo assembly of small genomes too. Exome, on the fly Bayesian SNP calculation, alignment directly to dbSNP, reference and OMIM, SNP report with browser. Demo of SeqMan NGEN, GUI or command line, BAM or SeqMan output formats. Hu genome would need 8 Tb, exome about 1 Tb disk space. GUI requires browsing to input template file, input sequence files (compatible with ILMN, 454, …), select assembly options (gap size, etc in advanced options, min base quality scores, etc.). Then click “assemble”. Visualization of BAM files in Lasergene software, not very RAM intensive. SNP report, filterable, annotated SNP, novel SNP, protein change, read depth. cDNA aligned to hu reference. Arraystar for RNA-seq and variant analysis - input is BAM or SNP report; excel like; gene level table, impact on protein (gene disruption), seemed easy to set and navigate. De novo assembler, ‘strategy view”, depth, pair consistency graphs on seqman. < $10k full suite, perpetual license + computer.  

Elaine Mardis. Wash U. Genome evolution, liquid tumors, collaboration with Tim Ley. Early 1970’s, Janet Rowley, leukemia under microscope, chromosomal changes. Correlative cancer genomics. Deep coverage as tumor block represents collective genomes of all nuclei. Deep sequencing gives a proportion of cells carrying each somatic mutation. In house process – SNP genotyping followed by sequencing. 30x fold PE ILMN reads, detect SNV, detect anomalous read pair mapping. Use custom capture probes for SNV and indels Tier 1-3 (coding, conserved regions) à to resequence and validate. Resequence at 1000x covg to get estimates of mutation rates or sample heterogeneity levels. AML1, primary tumor, relapse tumor, and normal skin. Plots of variant frequencies of relapse versus original tumor showing change in mutational frequencies (how many reads show a SNP). Cancers undergo clonal evolution, new mutatins induced by chemotherapy. Myelodysplastic syndromes (MDS) progress into AML. The MDS sample purified by flow sorting and sequenced, compared with normal skin and the secondary AML formed later in life. Mutations present both in MDS and AML, new mutation in AML. Newer mutations in a popn will occur at a lower frequency. The sec AML are highly heterogenous, therefore difficult to treat.  

Sam Aparicio. BCCA. saparcicio@bccrc.ca. Breast cancers. Genome aberrations. DNA level changes, ploidy to single mutations and small indels. RNA level changes. Part 1, > 1000 breast cancers, METABRIC consortia. Copy number changes in over 70% of the genome, filter by impact on gene expression (both cis and trans genes). Whole genome amplification and deletion landscapes, approx. 25 “hotspot” regions in the 1000 breast cancer genomes. Could be grouped into 10 patient types based on CNV grouping. Correlates with prognosis group. Group with almost no CNVs were 17% of population. Graph of gene expression against copy number variation, scales were chromosomal corordinates, showed trans effect of CNV on gene expression. Part 2, Triple negative breast cancer. Expression based clustering suggests at least 2 main types. 104 such cancers. Changes in alternative splicing reflecting the cell of origin rather than the cancer. Surprising diversity in these cancers. Less than 5 had BRCA mutations and did not feature strongly in expression changes.  

Chuck Perou. Univ N Carolina Chapel Hill. Mouse and human mammary tumors comparative genomics. PAM50 sub-type (50 genes by qPCR) from fresh or FFPE samples correlates with survival. Prognostic classification score. Mouse model with BRCA1 and TP53 (K14 Cre). Good test for chemotherapeutics. Mouse model did not show hormone responsive tumors. Most slides were gene exp array heat maps. aCGH on TP53- mouse and human tumors. Move into full genome sequencing on human samples with primary and metastatic tumors. Mouse tumor genomes also sequenced. Went through specific genes that were mutated.  

Peter Campbell, Sanger. Myeloid malignancies. Myelodysplastic syndromes, hard to diagnose (unexplained anemia), clonal stem cell disease, acquiring somatic mutations, recurrent gene mutations though infrequently mutated (< 5% for most). Complicated MDS classification. MDS exome sequencing versus germline by 454. 64 somatic mutations in 8, none in 1 patient. Fewer somatic mutations in exomes, excess of nonsynonymous mutations, 6 patients had mutation in SF3B1, 3 in DNMT3A, and 1 in TET2. Follow-up on 2,165 samples. 20% had mutations in SF3B1. But RARS subtype had > 67% of samples with mutations in SF3B1. SF3B1 also present in other cancers. Knockdown in zebrafish is lethal. Hematopoietic defect in zebrafish failing early transcriptional program. So took wbcs from SF3B1 patients of hematopoetic cells and did gene expression, many mitochondrial genes downregulated. Benign prognosis.

Adam Mumy, Bladder Cancer, 4th most common cancer. Prostate Stem cell antigen, gene expression, Tissue microarray, IHC with PSCA Ab. Risk allele highly expressed in tumor, non-risk allele not expressed. Henry Lin, Optical maps for genome assembly. Single DNA molecule, restriction digest, image analysis, restriction map. Very noisy 10% error plus 2Kb additive error (fragments within 2Kb may be combined). Applied de Bruijn graphs to create the restriction map. Tested on bacterial genomes.  

Desmond (?) Radiation hybrid mapping. Radiated mammalian à chromosome fragments à fused with hamster cells, some cells carry chromosome fragment. Couldn’t understand.  

Mark Boguski, Harvard Med. Clinical genomes becoming routine? Medical education, Hlth care information technology, regulation (FDA diagnostic tests approval), technology and tools, insurance coverage and reimbursement (CPT code, CNS – how much per test), genetic privacy and legal protections. Whole genome analysis as a universal diagnostic. First app as medical microbiology/metagenome. Formalin conventional fixative but not the best for molecular biology. Mentions David Bentley who presented story about an adenocarcinoma, 78 y/o male with throat discomfort, 4 months later, EGFR positive, given anti-EGFR, no effect. Did WGS and transcriptome. EGFR not expressed! WGS would needed to be multiple times during the course of the disease. $4k genome, $2k for data annotation, but urgently needed: a clinical grade knowledge base. Very few medical geneticists and genetic counselors, but lot’s of pathologists. There is a CPT code for molecular diagnostics. $18 per test! The innovators prescription..

Adding a scale to tools ...

Would be neat to add an inch and mm scale to adjustable tools such as adjustable wrenches, would give the option of setting up the wrench quickly and/or determining the size of objects.

Homemade camera copy stand

Just finished making a neat camera copy stand. Gives a broad range of camera position adjustments and is built surprisingly easily. This is for use with my electrophoresis and fluorescence experiments. Instructions for making it are in the attached PDF.

Filemaker pubmed database

Well am defintely learning more about connecting pubmed and filemaker. Have figured out how to import pubmed data dircetly into fmpro fields (though still cannot get the authors in) plus am using the webviewer to show the PDF and the pubmed site. If anyone out there can help me with converting xml data to fmpro, basically how to collapse multiple author fields in xml into one ...
Anyway have posted a screen shot (click on the thumbnail for a larger copy). If you need a copy of the database, please let me know.

Open source thermocycler - hair dryer version

Another potentially simpler approach is to use a travel hair dryer to circulate hot or cool air over the microcentrifuge tubes. The tubes would be arranged in a circle to get even heating from the hot air. A novel approach would be to heat the lid from the same hot air. Assuming that the larger the area exposed to the hot hair the higher the temperature it would be possible to design an integral heated lid that would maintain a temperature just above that of the tubes. To cool, the fan would be kept on but the heater would be turned off ... seems simple in concept but of course realizing a working deisgn will not be trivial. Would be good to create a thermal/fluid model of this to see if it will really work in practise. And/or protocotype this by using tin cans with holes for the tubes to see how long the heating and cooling cycle would take.